{
  "title": "Published overview",
  "models": [
    {
      "id": "auroratts",
      "name": "AuroraTTS-100K",
      "parameters_m": 0.09559999999999999,
      "approximate": false,
      "size_label": "95.6K",
      "size_note": "95.6K",
      "utmos": 3.085425,
      "wer": 0.1158,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/AuroraAI-Research/AuroraTTS-100K",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "auroratts_cpu_p1.wav",
          "model": "auroratts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "2.7395",
          "wer": "0.1429",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auroratts_cpu_p2.wav",
          "model": "auroratts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.2025",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auroratts_cpu_p3.wav",
          "model": "auroratts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.2034",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auroratts_cpu_p4.wav",
          "model": "auroratts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.1963",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "sanotts_amy",
      "name": "sanoTTS Amy",
      "parameters_m": 1.46,
      "approximate": false,
      "size_label": "1.46M",
      "size_note": "1.46M",
      "utmos": 3.8205999999999998,
      "wer": 0.066925,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/ampixa/sanoTTS",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "sanotts_amy_cpu_p1.wav",
          "model": "sanotts_amy",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.1379",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sanotts_amy_cpu_p2.wav",
          "model": "sanotts_amy",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.7211",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sanotts_amy_cpu_p3.wav",
          "model": "sanotts_amy",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.7677",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sanotts_amy_cpu_p4.wav",
          "model": "sanotts_amy",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.6557",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "smollertts",
      "name": "SmollerTTS",
      "parameters_m": 2.808792,
      "size_label": "2.81M",
      "size_note": "1,882,807 acoustic + 925,985 vocoder",
      "utmos": 3.553914248943329,
      "per_prompt": [
        {
          "index": 10,
          "status": "ok",
          "voice": "smollertts",
          "text": "Open the browser and read my email.",
          "path": "benchmark-1.wav",
          "sha256": "fc8e12cd6e72f001f4fa43a1dc6100b578be41a42aaabe4002093d4dcecb07f1",
          "kind": "benchmark",
          "prompt_id": 1,
          "type": "complete",
          "id": 11,
          "milliseconds": 451.40000000037253,
          "firstChunkMs": 451.30000000074506,
          "samples": 54272,
          "peak": 0.4160328507423401,
          "cacheHits": 0,
          "chunks": 1,
          "durations": [
            [
              11,
              12,
              9,
              5,
              6,
              3,
              3,
              8,
              4,
              19,
              7,
              10,
              4,
              12,
              7,
              13,
              7,
              7,
              15,
              13,
              9,
              10,
              14,
              4
            ]
          ],
          "controls": {
            "speed": 1,
            "pitch": 0,
            "energy": 0
          },
          "stageTimings": [
            {
              "encoder": 6.2000000001862645,
              "frames": 20.299999999813735,
              "vocoder": 422.5
            }
          ],
          "elapsedMs": 455.90000000037253,
          "firstReceivedMs": 451.79999999981374,
          "pronunciationNotes": [],
          "normalizedText": "Open the browser and read my email.",
          "utmos": 3.4720206260681152,
          "wav_sha256": "fc8e12cd6e72f001f4fa43a1dc6100b578be41a42aaabe4002093d4dcecb07f1",
          "wer": 0.0,
          "asr_transcript": "open the browser and read my email."
        },
        {
          "index": 11,
          "status": "ok",
          "voice": "smollertts",
          "text": "I'll start a new git branch, push the changes, and open a pull request when the tests pass.",
          "path": "benchmark-2.wav",
          "sha256": "1323fb5f6030ef3c45055cb16fbb6fde9e4eea51c046f63c38b403c39b98f311",
          "kind": "benchmark",
          "prompt_id": 2,
          "type": "complete",
          "id": 12,
          "milliseconds": 1212.4000000003725,
          "firstChunkMs": 1212.0999999996275,
          "samples": 142592,
          "peak": 0.48737382888793945,
          "cacheHits": 0,
          "chunks": 1,
          "durations": [
            [
              9,
              14,
              4,
              15,
              5,
              11,
              6,
              7,
              4,
              13,
              8,
              6,
              7,
              12,
              5,
              6,
              16,
              4,
              17,
              19,
              6,
              4,
              15,
              4,
              3,
              14,
              13,
              5,
              5,
              13,
              16,
              23,
              8,
              9,
              13,
              7,
              4,
              4,
              4,
              15,
              1,
              11,
              7,
              3,
              12,
              10,
              8,
              7,
              4,
              5,
              5,
              3,
              3,
              12,
              9,
              10,
              7,
              5,
              11,
              26,
              20,
              5
            ]
          ],
          "controls": {
            "speed": 1,
            "pitch": 0,
            "energy": 0
          },
          "stageTimings": [
            {
              "encoder": 12,
              "frames": 49.200000000186265,
              "vocoder": 1145.7999999998137
            }
          ],
          "elapsedMs": 1216.7999999998137,
          "firstReceivedMs": 1212.6000000005588,
          "pronunciationNotes": [],
          "normalizedText": "I'll start a new git branch, push the changes, and open a pull request when the tests pass.",
          "utmos": 4.107442378997803,
          "wav_sha256": "1323fb5f6030ef3c45055cb16fbb6fde9e4eea51c046f63c38b403c39b98f311",
          "wer": 0.0,
          "asr_transcript": "I'll start a new git branch, push the changes, and open a pull request when the tests pass."
        },
        {
          "index": 12,
          "status": "ok",
          "voice": "smollertts",
          "text": "The Parakeet TDT zero point six billion parameter model achieves one point six nine percent word error rate on LibriSpeech test-clean, beating Whisper Large V3 at two point seven percent while running at over two thousand times realtime on a single GPU.",
          "path": "benchmark-3.wav",
          "sha256": "cc15f7c51cc812e11a89663fe41a51c680bdc904936d5f21f65838db98468af6",
          "kind": "benchmark",
          "prompt_id": 3,
          "type": "complete",
          "id": 13,
          "milliseconds": 2426.5,
          "firstChunkMs": 2426.100000000559,
          "samples": 372480,
          "peak": 0.39931631088256836,
          "cacheHits": 0,
          "chunks": 1,
          "durations": [
            [
              15,
              4,
              3,
              11,
              6,
              7,
              5,
              12,
              10,
              6,
              9,
              9,
              8,
              13,
              10,
              14,
              8,
              8,
              8,
              6,
              12,
              12,
              4,
              11,
              5,
              7,
              5,
              9,
              5,
              10,
              4,
              5,
              9,
              4,
              7,
              10,
              5,
              4,
              6,
              6,
              10,
              10,
              2,
              3,
              8,
              5,
              15,
              11,
              4,
              7,
              6,
              8,
              7,
              10,
              11,
              4,
              10,
              5,
              7,
              8,
              6,
              16,
              4,
              8,
              4,
              11,
              6,
              8,
              5,
              10,
              7,
              9,
              12,
              5,
              12,
              11,
              7,
              9,
              4,
              11,
              5,
              7,
              3,
              4,
              10,
              6,
              9,
              12,
              7,
              6,
              7,
              9,
              10,
              4,
              14,
              12,
              22,
              5,
              8,
              5,
              5,
              7,
              6,
              4,
              8,
              5,
              6,
              9,
              10,
              6,
              10,
              4,
              9,
              6,
              15,
              5,
              8,
              9,
              8,
              12,
              16,
              7,
              13,
              7,
              4,
              5,
              5,
              8,
              6,
              11,
              5,
              8,
              5,
              14,
              7,
              9,
              7,
              5,
              6,
              6,
              9,
              5,
              12,
              6,
              5,
              12,
              13,
              7,
              16,
              7,
              5,
              5,
              9,
              18,
              7,
              9,
              9,
              8,
              8,
              9,
              15,
              7,
              8,
              4,
              5,
              15,
              9,
              4,
              4,
              12,
              13,
              15,
              14,
              9,
              14,
              19,
              10
            ]
          ],
          "controls": {
            "speed": 1,
            "pitch": 0,
            "energy": 0
          },
          "stageTimings": [
            {
              "encoder": 34.299999999813735,
              "frames": 135.8999999994412,
              "vocoder": 2241.4000000003725
            }
          ],
          "elapsedMs": 2432.7999999998137,
          "firstReceivedMs": 2426.5999999996275,
          "pronunciationNotes": [
            {
              "word": "TDT",
              "method": "spelled",
              "reading": "T D T"
            },
            {
              "word": "LibriSpeech",
              "method": "estimated",
              "ipa": "ˈɫɪbɹɪspitʃ"
            },
            {
              "word": "Vthree",
              "method": "estimated",
              "ipa": "ˈvθɹi"
            }
          ],
          "normalizedText": "The Parakeet TDT zero point six billion parameter model achieves one point six nine percent word error rate on LibriSpeech test-clean, beating Whisper Large Vthree at two point seven percent while running at over two thousand times realtime on a single G P U.",
          "utmos": 2.6406402587890625,
          "wav_sha256": "cc15f7c51cc812e11a89663fe41a51c680bdc904936d5f21f65838db98468af6",
          "wer": 0.21212121212121213,
          "asr_transcript": "the parakeet tdt zero point six billion parameter model achieves one point six nine per cent word error rate on a libre speech test clean beating whisper large free at two point seven per cent while running at over two thousand times real time on a single gpu"
        },
        {
          "index": 13,
          "status": "ok",
          "voice": "smollertts",
          "text": "Run pytest tests slash test underscore voice dot py with verbose flag and capture flag set to no.",
          "path": "benchmark-4.wav",
          "sha256": "0ee9de4cb64fdf5e4e6cf1a87da9ba929c7e0047a8636825c8da28873fe1ea39",
          "kind": "benchmark",
          "prompt_id": 4,
          "type": "complete",
          "id": 14,
          "milliseconds": 721.1999999992549,
          "firstChunkMs": 721.0999999996275,
          "samples": 150272,
          "peak": 0.4373336732387543,
          "cacheHits": 0,
          "chunks": 1,
          "durations": [
            [
              5,
              5,
              9,
              9,
              6,
              11,
              5,
              6,
              9,
              7,
              7,
              10,
              10,
              7,
              6,
              10,
              5,
              13,
              11,
              8,
              8,
              8,
              5,
              7,
              6,
              5,
              8,
              10,
              7,
              7,
              3,
              7,
              16,
              10,
              6,
              12,
              10,
              6,
              13,
              6,
              4,
              4,
              6,
              8,
              9,
              14,
              9,
              9,
              4,
              18,
              7,
              13,
              6,
              10,
              9,
              5,
              7,
              8,
              11,
              5,
              16,
              6,
              12,
              7,
              7,
              7,
              4,
              8,
              30,
              5
            ]
          ],
          "controls": {
            "speed": 1,
            "pitch": 0,
            "energy": 0
          },
          "stageTimings": [
            {
              "encoder": 8,
              "frames": 30.399999999441206,
              "vocoder": 678.2000000001863
            }
          ],
          "elapsedMs": 724.7000000001863,
          "firstReceivedMs": 721.3999999994412,
          "pronunciationNotes": [
            {
              "word": "pytest",
              "method": "estimated",
              "ipa": "ˈpaɪtɛst"
            },
            {
              "word": "underscore",
              "method": "estimated",
              "ipa": "ˌʌndɝˈskɔɹ"
            }
          ],
          "normalizedText": "Run pytest tests slash test underscore voice dot py with verbose flag and capture flag set to no.",
          "utmos": 3.995553731918335,
          "wav_sha256": "0ee9de4cb64fdf5e4e6cf1a87da9ba929c7e0047a8636825c8da28873fe1ea39",
          "wer": 0.05555555555555555,
          "asr_transcript": "Run pytest test slash test underscore voice dot py with verbose flag and capture flag set to no."
        }
      ],
      "source": "https://github.com/Ehdunhackme/SmollerTTS",
      "local": true,
      "judge": {
        "repository": "https://github.com/tarepan/SpeechMOS",
        "revision": "ed25eacbfa42b99156c36ebec67a733b5dbb9b79",
        "checkpoint_sha256": "38aa51ab79e2a4e09a1449758a4b37e9cbb2e8235a49662a732d33a9ba1e9bff",
        "torch": "2.6.0+cu124",
        "librosa": "0.11.0",
        "preprocessing": "librosa.load(sr=16000, mono=True), float32, no normalization; UTMOS22 strong, CPU eval"
      },
      "runtime": "Browser CPU",
      "n": 4,
      "wer": 0.06691919191919192,
      "wer_aggregation": "unweighted arithmetic mean of four prompt WERs, matching published overview",
      "wer_provenance": {
        "recognizer": "openai/whisper-large-v3",
        "revision": "06f233fe06e710322aca913c1bc4249a0d71fce1",
        "weight_sha256": "a8e94b85976e5864ba3e9525c7e6c83b2a1eca42d4b797a0c7c24d778e40fd95",
        "scorer_source_sha256": "7afd506d94cd629d5f017320b5e4609b9deff68e5ada65798cd58a691052232c",
        "source_revision": "1dd3a81a50dd21fb2f4bce89df27eada996ba57f",
        "device": "CPU",
        "dtype": "float32",
        "precision_note": "CPU FP32; the checkpoint is pinned. Published peer checkpoint and package revisions are not all pinned.",
        "threads": 2,
        "preprocessing": "librosa 16kHz mono float32, full waveform; WhisperProcessor default feature extraction then features cast to model FP32; no trimming or loudness normalization",
        "generation": {
          "language": "english",
          "task": "transcribe",
          "other_options": "pinned generation_config defaults"
        },
        "normalization": "Whisper tokenizer English normalizer, as in pinned tts-bench scorer; no custom post-hoc fixes",
        "aggregation": "unweighted arithmetic mean of four prompt WERs, matching published overview",
        "versions": {
          "torch": "2.6.0+cu124",
          "transformers": "5.13.1",
          "librosa": "0.11.0",
          "jiwer": "4.0.0",
          "huggingface_hub": "1.23.0"
        },
        "n": 4,
        "report": "./smollertts-wer.json"
      },
      "render_settings": {
        "speed": 1,
        "pitch": 0,
        "energy": 0,
        "punctuation": false,
        "note": "Frozen overview recordings; optional punctuation steering is disabled for reproduction."
      }
    },
    {
      "id": "inflect_nano",
      "name": "Inflect-Nano v2",
      "parameters_m": 3.96,
      "approximate": false,
      "size_label": "3.96M",
      "size_note": "3.96M",
      "utmos": 4.4332,
      "wer": 0.053025,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/owensong/Inflect-Nano-v2",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "inflect_nano_cpu_p1.wav",
          "model": "inflect_nano",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4638",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "inflect_nano_cpu_p2.wav",
          "model": "inflect_nano",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4645",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "inflect_nano_cpu_p3.wav",
          "model": "inflect_nano",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3446",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "inflect_nano_cpu_p4.wav",
          "model": "inflect_nano",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4599",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "tinyvoca",
      "name": "tiny-voca",
      "parameters_m": 4.5,
      "approximate": false,
      "size_label": "4.5M",
      "size_note": "4.5M",
      "utmos": 4.403675,
      "wer": 0.06565,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/n4ze3m/tiny-voca",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "tinyvoca_cpu_p1.wav",
          "model": "tinyvoca",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3633",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "tinyvoca_cpu_p2.wav",
          "model": "tinyvoca",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4631",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "tinyvoca_cpu_p3.wav",
          "model": "tinyvoca",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3803",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "tinyvoca_cpu_p4.wav",
          "model": "tinyvoca",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4080",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "vaniq",
      "name": "Vaniq-Edge",
      "parameters_m": 8.91,
      "approximate": false,
      "size_label": "8.91M",
      "size_note": "8.91M",
      "utmos": 4.272325,
      "wer": 0.09342500000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Abiray/Vaniq-Edge",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "vaniq_cpu_p1.wav",
          "model": "vaniq",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3223",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vaniq_cpu_p2.wav",
          "model": "vaniq",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3753",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vaniq_cpu_p3.wav",
          "model": "vaniq",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.2767",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vaniq_cpu_p4.wav",
          "model": "vaniq",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.1150",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "inflect_micro",
      "name": "Inflect-Micro v2",
      "parameters_m": 9.36,
      "approximate": false,
      "size_label": "9.36M",
      "size_note": "9.36M",
      "utmos": 4.444625,
      "wer": 0.0303,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/owensong/Inflect-Micro-v2",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "inflect_micro_cpu_p1.wav",
          "model": "inflect_micro",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3869",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "inflect_micro_cpu_p2.wav",
          "model": "inflect_micro",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4371",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "inflect_micro_cpu_p3.wav",
          "model": "inflect_micro",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4897",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "inflect_micro_cpu_p4.wav",
          "model": "inflect_micro",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4648",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "kittentts08",
      "name": "KittenTTS Nano 0.8",
      "parameters_m": 15.0,
      "approximate": false,
      "size_label": "15M",
      "size_note": "15M",
      "utmos": 4.2857,
      "wer": 0.0442,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/KittenML/kitten-tts-nano-0.8",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "kittentts08_cpu_p1.wav",
          "model": "kittentts08",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.1598",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts08_cpu_p2.wav",
          "model": "kittentts08",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4692",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts08_cpu_p3.wav",
          "model": "kittentts08",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0874",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts08_cpu_p4.wav",
          "model": "kittentts08",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4264",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "piper",
      "name": "Piper",
      "parameters_m": 15.0,
      "approximate": true,
      "size_label": "~15M",
      "size_note": "~15M",
      "utmos": 4.31135,
      "wer": 0.0442,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://github.com/OHF-Voice/piper1-gpl",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "piper_cpu_p1.wav",
          "model": "piper",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3353",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "piper_cpu_p2.wav",
          "model": "piper",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4154",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "piper_cpu_p3.wav",
          "model": "piper",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0490",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "piper_cpu_p4.wav",
          "model": "piper",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4457",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "steertts_noflow",
      "name": "SteerTTS (no refiner)",
      "parameters_m": 23.1,
      "approximate": false,
      "size_label": "23.1M",
      "size_note": "23.1M",
      "utmos": 3.95445,
      "wer": 0.064925,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/WhissleAI/steer-tts",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "steertts_noflow_cpu_p1.wav",
          "model": "steertts_noflow",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.7814",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "steertts_noflow_cpu_p2.wav",
          "model": "steertts_noflow",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3457",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "steertts_noflow_cpu_p3.wav",
          "model": "steertts_noflow",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.7720",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "steertts_noflow_cpu_p4.wav",
          "model": "steertts_noflow",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.9187",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "woof",
      "name": "Underdog Woof",
      "parameters_m": 38.2,
      "approximate": false,
      "size_label": "38.2M",
      "size_note": "38.2M (23.0M LM + 15.3M decoder)",
      "utmos": 4.445,
      "wer": 0.08965,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/ConwayResearch/Underdog-Woof",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "woof_cpu_p1.wav",
          "model": "woof",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.2935",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "woof_cpu_p2.wav",
          "model": "woof",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4858",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "woof_cpu_p3.wav",
          "model": "woof",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5055",
          "wer": "0.3030",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "woof_cpu_p4.wav",
          "model": "woof",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4952",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "steertts",
      "name": "SteerTTS",
      "parameters_m": 40.8,
      "approximate": false,
      "size_label": "40.8M",
      "size_note": "40.8M",
      "utmos": 4.0767,
      "wer": 0.064925,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/WhissleAI/steer-tts",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "steertts_cpu_p1.wav",
          "model": "steertts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8180",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "steertts_cpu_p2.wav",
          "model": "steertts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4333",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "steertts_cpu_p3.wav",
          "model": "steertts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.6707",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "steertts_cpu_p4.wav",
          "model": "steertts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3848",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "linnet",
      "name": "LinnetTTS",
      "parameters_m": 41.4,
      "approximate": false,
      "size_label": "41.4M",
      "size_note": "41.4M (31.5M LM + 9.9M decoder)",
      "utmos": 4.178075,
      "wer": 0.08837500000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/VoxRT/linnet-en-35m",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "linnet_cpu_p1.wav",
          "model": "linnet",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.9120",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "linnet_cpu_p2.wav",
          "model": "linnet",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.0723",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "linnet_cpu_p3.wav",
          "model": "linnet",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.2736",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "linnet_cpu_p4.wav",
          "model": "linnet",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4544",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "supratts",
      "name": "SupraTTS 0.1 Beta",
      "parameters_m": 43.5,
      "approximate": false,
      "size_label": "43.5M",
      "size_note": "43.5M",
      "utmos": 3.3011749999999997,
      "wer": 0.058075,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/SupraLabs/SupraTTS-0.1-Beta",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "supratts_cpu_p1.wav",
          "model": "supratts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "2.8303",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "supratts_cpu_p2.wav",
          "model": "supratts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.4602",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "supratts_cpu_p3.wav",
          "model": "supratts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.3621",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "supratts_cpu_p4.wav",
          "model": "supratts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.5521",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "melotts",
      "name": "MeloTTS",
      "parameters_m": 52.0,
      "approximate": true,
      "size_label": "~52M",
      "size_note": "~52M",
      "utmos": 3.7640000000000002,
      "wer": 0.152025,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/myshell-ai/MeloTTS-English",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "melotts_cpu_p1.wav",
          "model": "melotts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.5850",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "melotts_cpu_p2.wav",
          "model": "melotts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.1920",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "melotts_cpu_p3.wav",
          "model": "melotts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.6638",
          "wer": "0.3333",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "melotts_cpu_p4.wav",
          "model": "melotts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.6152",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "ribt",
      "name": "ribt",
      "parameters_m": 54.0,
      "approximate": false,
      "size_label": "54M",
      "size_note": "54M (18.65M acoustic + 21.88M flow + 13.5M vocoder)",
      "utmos": 4.254825,
      "wer": 0.071975,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/municipal-robotics/ribt",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "ribt_cpu_p1.wav",
          "model": "ribt",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3675",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ribt_cpu_p2.wav",
          "model": "ribt",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4508",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ribt_cpu_p3.wav",
          "model": "ribt",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.1811",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ribt_cpu_p4.wav",
          "model": "ribt",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.0199",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "soprano",
      "name": "Soprano 1.1 80M",
      "parameters_m": 80.0,
      "approximate": false,
      "size_label": "80M",
      "size_note": "80M",
      "utmos": 4.032675,
      "wer": 0.051775,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/ekwek/Soprano-1.1-80M",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "soprano_cpu_p1.wav",
          "model": "soprano",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.2369",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "soprano_cpu_p2.wav",
          "model": "soprano",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2407",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "soprano_cpu_p3.wav",
          "model": "soprano",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.8958",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "soprano_cpu_p4.wav",
          "model": "soprano",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.7573",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "kokoro",
      "name": "Kokoro",
      "parameters_m": 82.0,
      "approximate": false,
      "size_label": "82M",
      "size_note": "82M",
      "utmos": 4.504925,
      "wer": 0.08080000000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/hexgrad/Kokoro-82M",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "kokoro_cpu_p1.wav",
          "model": "kokoro",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4325",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kokoro_cpu_p2.wav",
          "model": "kokoro",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.5182",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kokoro_cpu_p3.wav",
          "model": "kokoro",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5426",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kokoro_cpu_p4.wav",
          "model": "kokoro",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.5264",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "hojo_40m",
      "name": "Hojo-TTS-Light 40M",
      "parameters_m": 84.5,
      "approximate": false,
      "size_label": "84.5M",
      "size_note": "84.5M (40.4M LM + 13.3M fine + 30.8M decoder)",
      "utmos": 4.3623,
      "wer": 0.118675,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/HojoAI/Hojo-TTS-Light-40M",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "hojo_40m_cpu_p1.wav",
          "model": "hojo_40m",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.9049",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "hojo_40m_cpu_p2.wav",
          "model": "hojo_40m",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4878",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "hojo_40m_cpu_p3.wav",
          "model": "hojo_40m",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5397",
          "wer": "0.3636",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "hojo_40m_cpu_p4.wav",
          "model": "hojo_40m",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.5168",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "supertonic",
      "name": "Supertonic 3",
      "parameters_m": 99.0,
      "approximate": false,
      "size_label": "99M",
      "size_note": "99M",
      "utmos": 4.4632000000000005,
      "wer": 0.0745,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Supertone/supertonic-3",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "supertonic_cpu_p1.wav",
          "model": "supertonic",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4221",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "supertonic_cpu_p2.wav",
          "model": "supertonic",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4389",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "supertonic_cpu_p3.wav",
          "model": "supertonic",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5159",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "supertonic_cpu_p4.wav",
          "model": "supertonic",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4759",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "moss_tts_nano",
      "name": "MOSS-TTS-Nano",
      "parameters_m": 100.0,
      "approximate": false,
      "size_label": "100M",
      "size_note": "100M",
      "utmos": 4.098025,
      "wer": 0.05935,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/OpenMOSS-Team/MOSS-TTS-Nano",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "moss_tts_nano_cpu_p1.wav",
          "model": "moss_tts_nano",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.0145",
          "wer": "0.0000",
          "sim": "",
          "health": "gap"
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_nano_cpu_p2.wav",
          "model": "moss_tts_nano",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2648",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_nano_cpu_p3.wav",
          "model": "moss_tts_nano",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0897",
          "wer": "0.1818",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_nano_cpu_p4.wav",
          "model": "moss_tts_nano",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.0231",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "openvoice",
      "name": "OpenVoice v2",
      "parameters_m": 100.0,
      "approximate": true,
      "size_label": "~100M",
      "size_note": "~100M",
      "utmos": 4.1025,
      "wer": 0.13725,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/myshell-ai/OpenVoiceV2",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "openvoice_cpu_p1.wav",
          "model": "openvoice",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.1181",
          "wer": "0.1429",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "openvoice_cpu_p2.wav",
          "model": "openvoice",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2812",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "openvoice_cpu_p3.wav",
          "model": "openvoice",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.1417",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "openvoice_cpu_p4.wav",
          "model": "openvoice",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.8690",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "pocket",
      "name": "Pocket-TTS v4",
      "parameters_m": 100.0,
      "approximate": false,
      "size_label": "100M",
      "size_note": "100M",
      "utmos": 4.32515,
      "wer": 0.053025,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://github.com/kyutai-labs/pocket-tts",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "pocket_cpu_p1.wav",
          "model": "pocket",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3870",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "pocket_cpu_p2.wav",
          "model": "pocket",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4290",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "pocket_cpu_p3.wav",
          "model": "pocket",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4457",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "pocket_cpu_p4.wav",
          "model": "pocket",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.0389",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "sopro",
      "name": "Sopro V2 Turbo",
      "parameters_m": 120.0,
      "approximate": false,
      "size_label": "120M",
      "size_note": "120M",
      "utmos": 4.00075,
      "wer": 0.08007500000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/samuel-vitorino/sopro-v2-turbo",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "sopro_cpu_p1.wav",
          "model": "sopro",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8414",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sopro_cpu_p2.wav",
          "model": "sopro",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.1550",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sopro_cpu_p3.wav",
          "model": "sopro",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.1992",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sopro_cpu_p4.wav",
          "model": "sopro",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.8074",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "sopro_streaming",
      "name": "Sopro V2 Turbo (streaming)",
      "parameters_m": 120.0,
      "approximate": false,
      "size_label": "120M",
      "size_note": "120M",
      "utmos": 3.637925,
      "wer": 0.0745,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/samuel-vitorino/sopro-v2-turbo",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "sopro_streaming_cpu_p1.wav",
          "model": "sopro_streaming",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.4262",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sopro_streaming_cpu_p2.wav",
          "model": "sopro_streaming",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.8630",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sopro_streaming_cpu_p3.wav",
          "model": "sopro_streaming",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.9167",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sopro_streaming_cpu_p4.wav",
          "model": "sopro_streaming",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.3458",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "luxtts",
      "name": "LuxTTS",
      "parameters_m": 123.0,
      "approximate": false,
      "size_label": "123M",
      "size_note": "123M",
      "utmos": 2.303125,
      "wer": 0.611025,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://github.com/ysharma3501/LuxTTS",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "luxtts_cpu_p1.wav",
          "model": "luxtts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "1.4784",
          "wer": "1.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "luxtts_cpu_p2.wav",
          "model": "luxtts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "2.0887",
          "wer": "0.3684",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "luxtts_cpu_p3.wav",
          "model": "luxtts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.1240",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "luxtts_cpu_p4.wav",
          "model": "luxtts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "2.5214",
          "wer": "0.8333",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "zipvoice",
      "name": "ZipVoice",
      "parameters_m": 123.0,
      "approximate": false,
      "size_label": "123M",
      "size_note": "123M",
      "utmos": 4.008900000000001,
      "wer": 0.071975,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/k2-fsa/ZipVoice",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "zipvoice_cpu_p1.wav",
          "model": "zipvoice",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8373",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "zipvoice_cpu_p2.wav",
          "model": "zipvoice",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3268",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "zipvoice_cpu_p3.wav",
          "model": "zipvoice",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.2281",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "zipvoice_cpu_p4.wav",
          "model": "zipvoice",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.6434",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "scyllasband",
      "name": "Scylla's Band",
      "parameters_m": 131.0,
      "approximate": true,
      "size_label": "~131M",
      "size_note": "~131M",
      "utmos": 4.15975,
      "wer": 0.02905,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/spybyscript/scyllasband",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "scyllasband_cpu_p1.wav",
          "model": "scyllasband",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.9200",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "scyllasband_cpu_p2.wav",
          "model": "scyllasband",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3353",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "scyllasband_cpu_p3.wav",
          "model": "scyllasband",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0143",
          "wer": "0.0606",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "scyllasband_cpu_p4.wav",
          "model": "scyllasband",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3694",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "styletts2",
      "name": "StyleTTS 2",
      "parameters_m": 148.0,
      "approximate": true,
      "size_label": "~148M",
      "size_note": "~148M",
      "utmos": 4.3717749999999995,
      "wer": 0.176575,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://github.com/yl4579/StyleTTS2",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "styletts2_cpu_p1.wav",
          "model": "styletts2",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3054",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "styletts2_cpu_p2.wav",
          "model": "styletts2",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4536",
          "wer": "0.1053",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "styletts2_cpu_p3.wav",
          "model": "styletts2",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4350",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "styletts2_cpu_p4.wav",
          "model": "styletts2",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.2931",
          "wer": "0.3889",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "hojo_80m",
      "name": "Hojo-TTS-Light 80M",
      "parameters_m": 167.0,
      "approximate": false,
      "size_label": "167M",
      "size_note": "167M (89.4M LM + 20.3M codec enc + 45.0M dec + 12.0M spk)",
      "utmos": 4.3414,
      "wer": 0.066925,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/HojoAI/Hojo-TTS-Light",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "hojo_80m_cpu_p1.wav",
          "model": "hojo_80m",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.0738",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "hojo_80m_cpu_p2.wav",
          "model": "hojo_80m",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4999",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "hojo_80m_cpu_p3.wav",
          "model": "hojo_80m",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3094",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "hojo_80m_cpu_p4.wav",
          "model": "hojo_80m",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4825",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "audio8_01b",
      "name": "Audio8 TTS Preview 0.1B",
      "parameters_m": 170.0,
      "approximate": false,
      "size_label": "170M",
      "size_note": "170M",
      "utmos": 4.40105,
      "wer": 0.113275,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Audio8/Audio8-TTS-Preview-0.1b",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "audio8_01b_cpu_p1.wav",
          "model": "audio8_01b",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4641",
          "wer": "0.1429",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_01b_cpu_p2.wav",
          "model": "audio8_01b",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3816",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_01b_cpu_p3.wav",
          "model": "audio8_01b",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4993",
          "wer": "0.0909",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_01b_cpu_p4.wav",
          "model": "audio8_01b",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.2592",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "neutts_nano",
      "name": "NeuTTS Nano",
      "parameters_m": 229.0,
      "approximate": false,
      "size_label": "229M",
      "size_note": "229M",
      "utmos": 4.116474999999999,
      "wer": 0.272725,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/neuphonic/neutts-nano-q4-gguf",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "neutts_nano_cpu_p1.wav",
          "model": "neutts_nano",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.7968",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "neutts_nano_cpu_p2.wav",
          "model": "neutts_nano",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2645",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "neutts_nano_cpu_p3.wav",
          "model": "neutts_nano",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3744",
          "wer": "0.7576",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "neutts_nano_cpu_p4.wav",
          "model": "neutts_nano",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.0302",
          "wer": "0.3333",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "f5tts",
      "name": "F5-TTS v1",
      "parameters_m": 330.0,
      "approximate": false,
      "size_label": "330M",
      "size_note": "330M",
      "utmos": 4.02095,
      "wer": 0.21407500000000002,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/SWivid/F5-TTS",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "f5tts_cpu_p1.wav",
          "model": "f5tts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.9900",
          "wer": "0.1429",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "f5tts_cpu_p2.wav",
          "model": "f5tts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.1206",
          "wer": "0.1579",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "f5tts_cpu_p3.wav",
          "model": "f5tts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0163",
          "wer": "0.3333",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "f5tts_cpu_p4.wav",
          "model": "f5tts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.9569",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "magpie",
      "name": "Magpie-TTS",
      "parameters_m": 357.0,
      "approximate": false,
      "size_label": "357M",
      "size_note": "357M",
      "utmos": 4.38325,
      "wer": 0.07955,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/nvidia/magpie_tts_multilingual_357m",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "magpie_cpu_p1.wav",
          "model": "magpie",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3280",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "magpie_cpu_p2.wav",
          "model": "magpie",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3856",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "magpie_cpu_p3.wav",
          "model": "magpie",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4122",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "magpie_cpu_p4.wav",
          "model": "magpie",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4072",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "vibevoice",
      "name": "VibeVoice Realtime 0.5B",
      "parameters_m": 500.0,
      "approximate": false,
      "size_label": "0.5B",
      "size_note": "0.5B",
      "utmos": 4.21515,
      "wer": 0.101,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/microsoft/VibeVoice-Realtime-0.5B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "vibevoice_cpu_p1.wav",
          "model": "vibevoice",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.1683",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_cpu_p2.wav",
          "model": "vibevoice",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4197",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_cpu_p3.wav",
          "model": "vibevoice",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0974",
          "wer": "0.1818",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_cpu_p4.wav",
          "model": "vibevoice",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.1752",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "cosyvoice",
      "name": "CosyVoice 3 0.5B",
      "parameters_m": 500.0,
      "approximate": false,
      "size_label": "0.5B",
      "size_note": "0.5B",
      "utmos": 3.9748,
      "wer": 0.05935,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/FunAudioLLM/Fun-CosyVoice3-0.5B-2512",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "cosyvoice_cuda_p1.wav",
          "model": "cosyvoice",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.7331",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "cosyvoice_cuda_p2.wav",
          "model": "cosyvoice",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.8639",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "cosyvoice_cuda_p3.wav",
          "model": "cosyvoice",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.9281",
          "wer": "0.1818",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "cosyvoice_cuda_p4.wav",
          "model": "cosyvoice",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3741",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "fish_15",
      "name": "Fish Speech 1.5",
      "parameters_m": 500.0,
      "approximate": true,
      "size_label": "~500M",
      "size_note": "~500M",
      "utmos": 3.8691500000000003,
      "wer": 0.16465,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/fishaudio/fish-speech-1.5",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "fish_15_cpu_p1.wav",
          "model": "fish_15",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.0609",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "fish_15_cpu_p2.wav",
          "model": "fish_15",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.6815",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "fish_15_cpu_p3.wav",
          "model": "fish_15",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.8263",
          "wer": "0.2727",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "fish_15_cpu_p4.wav",
          "model": "fish_15",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.9079",
          "wer": "0.3333",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "miratts",
      "name": "MiraTTS",
      "parameters_m": 500.0,
      "approximate": false,
      "size_label": "0.5B",
      "size_note": "0.5B",
      "utmos": 3.803,
      "wer": 0.09975,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/YatharthS/MiraTTS",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "miratts_cuda_p1.wav",
          "model": "miratts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.5780",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "miratts_cuda_p2.wav",
          "model": "miratts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.8460",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "miratts_cuda_p3.wav",
          "model": "miratts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0782",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "miratts_cuda_p4.wav",
          "model": "miratts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.7098",
          "wer": "0.2778",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "qwentts_06b_custom",
      "name": "Qwen3-TTS 0.6B CustomVoice",
      "parameters_m": 600.0,
      "approximate": false,
      "size_label": "0.6B",
      "size_note": "0.6B",
      "utmos": 4.372450000000001,
      "wer": 0.087125,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "qwentts_06b_custom_cpu_p1.wav",
          "model": "qwentts_06b_custom",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.2107",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_06b_custom_cpu_p2.wav",
          "model": "qwentts_06b_custom",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3728",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_06b_custom_cpu_p3.wav",
          "model": "qwentts_06b_custom",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4724",
          "wer": "0.1818",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_06b_custom_cpu_p4.wav",
          "model": "qwentts_06b_custom",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4339",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "audio8_fast",
      "name": "Audio8 TTS 0.6B (compiled)",
      "parameters_m": 601.0,
      "approximate": false,
      "size_label": "601M",
      "size_note": "601M",
      "utmos": 4.4254750000000005,
      "wer": 0.058075,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/scrappylabsai/audio8-tts-fastpath",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "audio8_fast_cuda_p1.wav",
          "model": "audio8_fast",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3451",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_fast_cuda_p2.wav",
          "model": "audio8_fast",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4513",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_fast_cuda_p3.wav",
          "model": "audio8_fast",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4998",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_fast_cuda_p4.wav",
          "model": "audio8_fast",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4057",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "audio8",
      "name": "Audio8 TTS Preview 0.6B",
      "parameters_m": 601.0,
      "approximate": false,
      "size_label": "601M",
      "size_note": "601M",
      "utmos": 4.4285499999999995,
      "wer": 0.087125,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Audio8/Audio8-TTS-Preview-0.6b",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "audio8_cpu_p1.wav",
          "model": "audio8",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3567",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_cpu_p2.wav",
          "model": "audio8",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4186",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_cpu_p3.wav",
          "model": "audio8",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4959",
          "wer": "0.1818",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_cpu_p4.wav",
          "model": "audio8",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4430",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "audio8_realistic",
      "name": "Audio8 TTS Realistic 0.6B",
      "parameters_m": 601.0,
      "approximate": false,
      "size_label": "601M",
      "size_note": "601M",
      "utmos": 4.5062,
      "wer": 0.02905,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Edge0/Audio8-TTS-Realistic-0.6b",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "audio8_realistic_cpu_p1.wav",
          "model": "audio8_realistic",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4905",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_realistic_cpu_p2.wav",
          "model": "audio8_realistic",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.5252",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_realistic_cpu_p3.wav",
          "model": "audio8_realistic",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5124",
          "wer": "0.0606",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "audio8_realistic_cpu_p4.wav",
          "model": "audio8_realistic",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4967",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "chatterbox_turbo",
      "name": "ChatterBox Turbo",
      "parameters_m": 744.0,
      "approximate": false,
      "size_label": "744M",
      "size_note": "744M",
      "utmos": 4.26075,
      "wer": 0.0745,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/ResembleAI/chatterbox-turbo",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "chatterbox_turbo_cpu_p1.wav",
          "model": "chatterbox_turbo",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.0068",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "chatterbox_turbo_cpu_p2.wav",
          "model": "chatterbox_turbo",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4538",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "chatterbox_turbo_cpu_p3.wav",
          "model": "chatterbox_turbo",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.1424",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "chatterbox_turbo_cpu_p4.wav",
          "model": "chatterbox_turbo",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4400",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "neutts_air",
      "name": "NeuTTS Air",
      "parameters_m": 748.0,
      "approximate": false,
      "size_label": "748M",
      "size_note": "748M",
      "utmos": 3.78545,
      "wer": 0.138875,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/neuphonic/neutts-air",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "neutts_air_cpu_p1.wav",
          "model": "neutts_air",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.7506",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "neutts_air_cpu_p2.wav",
          "model": "neutts_air",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.7299",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "neutts_air_cpu_p3.wav",
          "model": "neutts_air",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0194",
          "wer": "0.3333",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "neutts_air_cpu_p4.wav",
          "model": "neutts_air",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.6419",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "coqui",
      "name": "Coqui XTTS-v2",
      "parameters_m": 750.0,
      "approximate": false,
      "size_label": "750M",
      "size_note": "750M",
      "utmos": 4.1856,
      "wer": 0.1127,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/coqui/XTTS-v2",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "coqui_cpu_p1.wav",
          "model": "coqui",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8860",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "coqui_cpu_p2.wav",
          "model": "coqui",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4112",
          "wer": "0.1579",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "coqui_cpu_p3.wav",
          "model": "coqui",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.2566",
          "wer": "0.1818",
          "sim": "",
          "health": "gap"
        },
        {
          "dir": "linux-default",
          "wav": "coqui_cpu_p4.wav",
          "model": "coqui",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.1886",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "zonos3_mini",
      "name": "ZONOS3-Mini (ggml Q4)",
      "parameters_m": 800.0,
      "approximate": false,
      "size_label": "800M",
      "size_note": "800M (MoE, ~100M active; Q4/Q8 GGUF)",
      "utmos": 3.932375,
      "wer": 0.08080000000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Zyphra/ZONOS3-Mini",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "zonos3_mini_cpu_p1.wav",
          "model": "zonos3_mini",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.5970",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "zonos3_mini_cpu_p2.wav",
          "model": "zonos3_mini",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.9750",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "zonos3_mini_cpu_p3.wav",
          "model": "zonos3_mini",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3312",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "zonos3_mini_cpu_p4.wav",
          "model": "zonos3_mini",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.8263",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "parler",
      "name": "Parler-TTS Mini v1",
      "parameters_m": 878.0,
      "approximate": false,
      "size_label": "878M",
      "size_note": "878M",
      "utmos": 3.5500249999999998,
      "wer": 0.1533,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/parler-tts/parler-tts-mini-v1",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "parler_cpu_p1.wav",
          "model": "parler",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.5439",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "parler_cpu_p2.wav",
          "model": "parler",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.5356",
          "wer": "0.0526",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "parler_cpu_p3.wav",
          "model": "parler",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.0563",
          "wer": "0.3939",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "parler_cpu_p4.wav",
          "model": "parler",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.0643",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "indextts25",
      "name": "IndexTTS-2.5",
      "parameters_m": 970.0,
      "approximate": false,
      "size_label": "0.97B",
      "size_note": "0.97B",
      "utmos": 3.768975,
      "wer": 0.15315,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/IndexTeam/IndexTTS-2.5",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "indextts25_cpu_p1.wav",
          "model": "indextts25",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.5103",
          "wer": "0.1429",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "indextts25_cpu_p2.wav",
          "model": "indextts25",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.7695",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "indextts25_cpu_p3.wav",
          "model": "indextts25",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.8551",
          "wer": "0.3030",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "indextts25_cpu_p4.wav",
          "model": "indextts25",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.9410",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "outetts",
      "name": "OuteTTS 1.0 1B",
      "parameters_m": 1000.0,
      "approximate": true,
      "size_label": "~1B",
      "size_note": "~1B",
      "utmos": 4.4265,
      "wer": 0.08585000000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/OuteAI/Llama-OuteTTS-1.0-1B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "outetts_cuda_p1.wav",
          "model": "outetts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4685",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "outetts_cuda_p2.wav",
          "model": "outetts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4144",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "outetts_cuda_p3.wav",
          "model": "outetts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4717",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "outetts_cuda_p4.wav",
          "model": "outetts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3514",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "omnivoice",
      "name": "OmniVoice",
      "parameters_m": 1000.0,
      "approximate": true,
      "size_label": "~1B",
      "size_note": "~1B",
      "utmos": 4.268975,
      "wer": 0.042925000000000005,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/k2-fsa/OmniVoice",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "omnivoice_cpu_p1.wav",
          "model": "omnivoice",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4385",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "omnivoice_cpu_p2.wav",
          "model": "omnivoice",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3471",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "omnivoice_cpu_p3.wav",
          "model": "omnivoice",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3990",
          "wer": "0.0606",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "omnivoice_cpu_p4.wav",
          "model": "omnivoice",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.8913",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "sesame",
      "name": "Sesame CSM-1B",
      "parameters_m": 1000.0,
      "approximate": false,
      "size_label": "1B",
      "size_note": "1B",
      "utmos": 4.18685,
      "wer": 0.19175,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/sesame/csm-1b",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "sesame_cpu_p1.wav",
          "model": "sesame",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.9686",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sesame_cpu_p2.wav",
          "model": "sesame",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.1180",
          "wer": "0.1053",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sesame_cpu_p3.wav",
          "model": "sesame",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.2514",
          "wer": "0.6061",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "sesame_cpu_p4.wav",
          "model": "sesame",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4094",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "chatterbox",
      "name": "ChatterBox",
      "parameters_m": 1200.0,
      "approximate": false,
      "size_label": "1.2B",
      "size_note": "1.2B",
      "utmos": 4.438725,
      "wer": 0.0505,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/ResembleAI/chatterbox",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "chatterbox_cpu_p1.wav",
          "model": "chatterbox",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4098",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "chatterbox_cpu_p2.wav",
          "model": "chatterbox",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4091",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "chatterbox_cpu_p3.wav",
          "model": "chatterbox",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4992",
          "wer": "0.0909",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "chatterbox_cpu_p4.wav",
          "model": "chatterbox",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4368",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "mars5",
      "name": "Mars5-TTS",
      "parameters_m": 1200.0,
      "approximate": false,
      "size_label": "1.2B",
      "size_note": "1.2B",
      "utmos": 3.52955,
      "wer": 0.34125,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/Camb-ai/mars5-tts",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "mars5_cuda_p1.wav",
          "model": "mars5",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.9009",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "mars5_cuda_p2.wav",
          "model": "mars5",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3533",
          "wer": "0.1579",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "mars5_cuda_p3.wav",
          "model": "mars5",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "1.5930",
          "wer": "0.8182",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "mars5_cuda_p4.wav",
          "model": "mars5",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.2710",
          "wer": "0.3889",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "metavoice",
      "name": "MetaVoice-1B",
      "parameters_m": 1200.0,
      "approximate": false,
      "size_label": "1.2B",
      "size_note": "1.2B",
      "utmos": 3.79555,
      "wer": 0.33725,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/metavoiceio/metavoice-1B-v0.1",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "metavoice_cuda_p1.wav",
          "model": "metavoice",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.6144",
          "wer": "0.2857",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "metavoice_cuda_p2.wav",
          "model": "metavoice",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.7250",
          "wer": "0.3158",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "metavoice_cuda_p3.wav",
          "model": "metavoice",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0111",
          "wer": "0.6364",
          "sim": "",
          "health": "gap"
        },
        {
          "dir": "linux-default",
          "wav": "metavoice_cuda_p4.wav",
          "model": "metavoice",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.8317",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "longcat_1b",
      "name": "LongCat-AudioDiT 1B",
      "parameters_m": 1420.0,
      "approximate": false,
      "size_label": "1.42B",
      "size_note": "1.42B",
      "utmos": 4.0535499999999995,
      "wer": 0.09342500000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/meituan-longcat/LongCat-AudioDiT-1B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "longcat_1b_cuda_p1.wav",
          "model": "longcat_1b",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.4646",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "longcat_1b_cuda_p2.wav",
          "model": "longcat_1b",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.9753",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "longcat_1b_cuda_p3.wav",
          "model": "longcat_1b",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3970",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "longcat_1b_cuda_p4.wav",
          "model": "longcat_1b",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3773",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "ming",
      "name": "Ming-3.0-TTS 0.5B",
      "parameters_m": 1480.0,
      "approximate": false,
      "size_label": "1.48B",
      "size_note": "1.48B",
      "utmos": 4.038275,
      "wer": 0.073225,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/inclusionAI/Ming-3.0-tts-0.5B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "ming_cuda_p1.wav",
          "model": "ming",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4233",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ming_cuda_p2.wav",
          "model": "ming",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2494",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ming_cuda_p3.wav",
          "model": "ming",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.8584",
          "wer": "0.1818",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ming_cuda_p4.wav",
          "model": "ming",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.6220",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "lfm2_audio",
      "name": "LFM2.5-Audio 1.5B",
      "parameters_m": 1500.0,
      "approximate": false,
      "size_label": "1.5B",
      "size_note": "1.5B",
      "utmos": 4.367925,
      "wer": 0.0947,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "lfm2_audio_cpu_p1.wav",
          "model": "lfm2_audio",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.2467",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "lfm2_audio_cpu_p2.wav",
          "model": "lfm2_audio",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4225",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "lfm2_audio_cpu_p3.wav",
          "model": "lfm2_audio",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4603",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "lfm2_audio_cpu_p4.wav",
          "model": "lfm2_audio",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3422",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "indextts",
      "name": "IndexTTS-2",
      "parameters_m": 1500.0,
      "approximate": false,
      "size_label": "1.5B",
      "size_note": "1.5B",
      "utmos": 3.754575,
      "wer": 0.0947,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/IndexTeam/IndexTTS-2",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "indextts_cpu_p1.wav",
          "model": "indextts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.6888",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "indextts_cpu_p2.wav",
          "model": "indextts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.7228",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "indextts_cpu_p3.wav",
          "model": "indextts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0602",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "indextts_cpu_p4.wav",
          "model": "indextts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.5465",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "vibevoice_15b",
      "name": "VibeVoice 1.5B",
      "parameters_m": 1500.0,
      "approximate": false,
      "size_label": "1.5B",
      "size_note": "1.5B",
      "utmos": 3.7057,
      "wer": 0.168075,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/microsoft/VibeVoice-1.5B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "vibevoice_15b_cpu_p1.wav",
          "model": "vibevoice_15b",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.6124",
          "wer": "0.0000",
          "sim": "0.4822",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_15b_cpu_p2.wav",
          "model": "vibevoice_15b",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "2.7570",
          "wer": "0.2632",
          "sim": "0.4189",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_15b_cpu_p3.wav",
          "model": "vibevoice_15b",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.1351",
          "wer": "0.2424",
          "sim": "0.7442",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_15b_cpu_p4.wav",
          "model": "vibevoice_15b",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3183",
          "wer": "0.1667",
          "sim": "0.7636",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "auk_base",
      "name": "AuK",
      "parameters_m": 1530.0,
      "approximate": false,
      "size_label": "1.53B",
      "size_note": "1.53B",
      "utmos": 4.160275,
      "wer": 0.063125,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/tencent/AuK",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "auk_base_cuda_p1.wav",
          "model": "auk_base",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.9453",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auk_base_cuda_p2.wav",
          "model": "auk_base",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2577",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auk_base_cuda_p3.wav",
          "model": "auk_base",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.2045",
          "wer": "0.0303",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auk_base_cuda_p4.wav",
          "model": "auk_base",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.2336",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "auk_flash",
      "name": "AuK-Flash",
      "parameters_m": 1530.0,
      "approximate": false,
      "size_label": "1.53B",
      "size_note": "1.53B",
      "utmos": 4.2747,
      "wer": 0.056825,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/tencent/AuK-Flash",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "auk_flash_cuda_p1.wav",
          "model": "auk_flash",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.9827",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auk_flash_cuda_p2.wav",
          "model": "auk_flash",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3673",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auk_flash_cuda_p3.wav",
          "model": "auk_flash",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4303",
          "wer": "0.0606",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "auk_flash_cuda_p4.wav",
          "model": "auk_flash",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3185",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "dia",
      "name": "Dia 1.6B-0626",
      "parameters_m": 1600.0,
      "approximate": false,
      "size_label": "1.6B",
      "size_note": "1.6B",
      "utmos": 3.02715,
      "wer": 0.101525,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/nari-labs/Dia-1.6B-0626",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "dia_cuda_p1.wav",
          "model": "dia",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "2.2590",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "dia_cuda_p2.wav",
          "model": "dia",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "2.5041",
          "wer": "0.0526",
          "sim": "",
          "health": "gap"
        },
        {
          "dir": "linux-default",
          "wav": "dia_cuda_p3.wav",
          "model": "dia",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.8012",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "dia_cuda_p4.wav",
          "model": "dia",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.5443",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "zonos",
      "name": "Zonos v0.1",
      "parameters_m": 1600.0,
      "approximate": false,
      "size_label": "1.6B",
      "size_note": "1.6B",
      "utmos": 4.103125,
      "wer": 0.0745,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Zyphra/Zonos-v0.1-transformer",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "zonos_cpu_p1.wav",
          "model": "zonos",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.1991",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "zonos_cpu_p2.wav",
          "model": "zonos",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.0869",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "zonos_cpu_p3.wav",
          "model": "zonos",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.9078",
          "wer": "0.2424",
          "sim": "",
          "health": "gap"
        },
        {
          "dir": "linux-default",
          "wav": "zonos_cpu_p4.wav",
          "model": "zonos",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.2187",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "kittentts2",
      "name": "KittenTTS 2",
      "parameters_m": 1700.0,
      "approximate": false,
      "size_label": "1.7B",
      "size_note": "1.7B",
      "utmos": 4.5058,
      "wer": 0.1414,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/KittenML/kitten-tts-2",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "kittentts2_cpu_p1.wav",
          "model": "kittentts2",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4832",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts2_cpu_p2.wav",
          "model": "kittentts2",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.5312",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts2_cpu_p3.wav",
          "model": "kittentts2",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4853",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts2_cpu_p4.wav",
          "model": "kittentts2",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.5235",
          "wer": "0.4444",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "kittentts2_cpp",
      "name": "KittenTTS 2 (C++ CPU)",
      "parameters_m": 1700.0,
      "approximate": false,
      "size_label": "1.7B",
      "size_note": "1.7B",
      "utmos": 4.51105,
      "wer": 0.058075,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://github.com/KittenML/kitten-tts-2-cpp",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "kittentts2_cpp_cpu_p1.wav",
          "model": "kittentts2_cpp",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4502",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts2_cpp_cpu_p2.wav",
          "model": "kittentts2_cpp",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.5086",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts2_cpp_cpu_p3.wav",
          "model": "kittentts2_cpp",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5553",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "kittentts2_cpp_cpu_p4.wav",
          "model": "kittentts2_cpp",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.5301",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "qwentts",
      "name": "Qwen3-TTS 1.7B Base",
      "parameters_m": 1700.0,
      "approximate": false,
      "size_label": "1.7B",
      "size_note": "1.7B",
      "utmos": 4.10515,
      "wer": 0.08080000000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/Qwen/Qwen3-TTS-12Hz-1.7B-Base",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "qwentts_cpu_p1.wav",
          "model": "qwentts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8649",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_cpu_p2.wav",
          "model": "qwentts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2956",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_cpu_p3.wav",
          "model": "qwentts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3935",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_cpu_p4.wav",
          "model": "qwentts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.8666",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "qwentts_fast",
      "name": "Qwen3-TTS 1.7B (CUDA-graph)",
      "parameters_m": 1700.0,
      "approximate": false,
      "size_label": "1.7B",
      "size_note": "1.7B",
      "utmos": 4.3022,
      "wer": 0.134175,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/Qwen/Qwen3-TTS-12Hz-1.7B-Base",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "qwentts_fast_cuda_p1.wav",
          "model": "qwentts_fast",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.2891",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_fast_cuda_p2.wav",
          "model": "qwentts_fast",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2680",
          "wer": "0.1579",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_fast_cuda_p3.wav",
          "model": "qwentts_fast",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3549",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "qwentts_fast_cuda_p4.wav",
          "model": "qwentts_fast",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.2968",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "lfm3_audio",
      "name": "LFM3-Audio 2B",
      "parameters_m": 2000.0,
      "approximate": false,
      "size_label": "2.0B",
      "size_note": "2.0B",
      "utmos": 4.43565,
      "wer": 0.036625,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/LiquidAI/LFM3-Audio-2B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "lfm3_audio_cpu_p1.wav",
          "model": "lfm3_audio",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4556",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "lfm3_audio_cpu_p2.wav",
          "model": "lfm3_audio",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4739",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "lfm3_audio_cpu_p3.wav",
          "model": "lfm3_audio",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3632",
          "wer": "0.0909",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "lfm3_audio_cpu_p4.wav",
          "model": "lfm3_audio",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4499",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "dots_tts",
      "name": "dots.tts (soar)",
      "parameters_m": 2000.0,
      "approximate": false,
      "size_label": "2B",
      "size_note": "2B",
      "utmos": 3.65415,
      "wer": 0.073225,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/rednote-hilab/dots.tts-soar",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "dots_tts_cuda_p1.wav",
          "model": "dots_tts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.2030",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "dots_tts_cuda_p2.wav",
          "model": "dots_tts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.0627",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "dots_tts_cuda_p3.wav",
          "model": "dots_tts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.8071",
          "wer": "0.1818",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "dots_tts_cuda_p4.wav",
          "model": "dots_tts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.5438",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "voxcpm",
      "name": "VoxCPM2",
      "parameters_m": 2000.0,
      "approximate": false,
      "size_label": "2B",
      "size_note": "2B",
      "utmos": 4.029125,
      "wer": 0.042925000000000005,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CPU",
      "source": "https://huggingface.co/openbmb/VoxCPM2",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "voxcpm_cpu_p1.wav",
          "model": "voxcpm",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.5886",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "voxcpm_cpu_p2.wav",
          "model": "voxcpm",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3816",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "voxcpm_cpu_p3.wav",
          "model": "voxcpm",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4548",
          "wer": "0.0606",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "voxcpm_cpu_p4.wav",
          "model": "voxcpm",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.6915",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "firered3",
      "name": "FireRedTTS3",
      "parameters_m": 2120.0,
      "approximate": false,
      "size_label": "2.12B",
      "size_note": "2.12B",
      "utmos": 4.13225,
      "wer": 0.102275,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/FireRedTeam/FireRedTTS3",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "firered3_cuda_p1.wav",
          "model": "firered3",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8036",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "firered3_cuda_p2.wav",
          "model": "firered3",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2154",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "firered3_cuda_p3.wav",
          "model": "firered3",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3135",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "firered3_cuda_p4.wav",
          "model": "firered3",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.1965",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "echo",
      "name": "Echo-TTS",
      "parameters_m": 2800.0,
      "approximate": true,
      "size_label": "~2.8B",
      "size_note": "~2.8B",
      "utmos": 4.211925,
      "wer": 0.0745,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/jordand/echo-tts-base",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "echo_cuda_p1.wav",
          "model": "echo",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.0375",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "echo_cuda_p2.wav",
          "model": "echo",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3128",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "echo_cuda_p3.wav",
          "model": "echo",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3427",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "echo_cuda_p4.wav",
          "model": "echo",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.1547",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "maya1",
      "name": "Maya1",
      "parameters_m": 3000.0,
      "approximate": false,
      "size_label": "3B",
      "size_note": "3B",
      "utmos": 4.4685749999999995,
      "wer": 0.0505,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/maya-research/maya1",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "maya1_cuda_p1.wav",
          "model": "maya1",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4692",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "maya1_cuda_p2.wav",
          "model": "maya1",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4683",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "maya1_cuda_p3.wav",
          "model": "maya1",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5015",
          "wer": "0.0909",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "maya1_cuda_p4.wav",
          "model": "maya1",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4353",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "orpheus",
      "name": "Orpheus TTS",
      "parameters_m": 3000.0,
      "approximate": false,
      "size_label": "3B",
      "size_note": "3B",
      "utmos": 4.0049,
      "wer": 0.102275,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/canopylabs/orpheus-3b-0.1-ft",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "orpheus_cuda_p1.wav",
          "model": "orpheus",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8070",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "orpheus_cuda_p2.wav",
          "model": "orpheus",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.9797",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "orpheus_cuda_p3.wav",
          "model": "orpheus",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0710",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "orpheus_cuda_p4.wav",
          "model": "orpheus",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.1619",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "step_editx",
      "name": "Step-Audio-EditX",
      "parameters_m": 3000.0,
      "approximate": false,
      "size_label": "3B",
      "size_note": "3B",
      "utmos": 4.2334000000000005,
      "wer": 0.10137499999999999,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/stepfun-ai/Step-Audio-EditX",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "step_editx_cuda_p1.wav",
          "model": "step_editx",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8416",
          "wer": "0.1429",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "step_editx_cuda_p2.wav",
          "model": "step_editx",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2529",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "step_editx_cuda_p3.wav",
          "model": "step_editx",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4287",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "step_editx_cuda_p4.wav",
          "model": "step_editx",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4104",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "dramabox",
      "name": "DramaBox",
      "parameters_m": 3300.0,
      "approximate": false,
      "size_label": "3.3B",
      "size_note": "3.3B",
      "utmos": 4.228949999999999,
      "wer": 0.09342500000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://github.com/resemble-ai/DramaBox",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "dramabox_cuda_p1.wav",
          "model": "dramabox",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.1697",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "dramabox_cuda_p2.wav",
          "model": "dramabox",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2881",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "dramabox_cuda_p3.wav",
          "model": "dramabox",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.0977",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "dramabox_cuda_p4.wav",
          "model": "dramabox",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3603",
          "wer": "0.2222",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "longcat_3p5b",
      "name": "LongCat-AudioDiT 3.5B",
      "parameters_m": 3830.0,
      "approximate": false,
      "size_label": "3.83B",
      "size_note": "3.83B",
      "utmos": 4.340725,
      "wer": 0.07955,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/meituan-longcat/LongCat-AudioDiT-3.5B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "longcat_3p5b_cuda_p1.wav",
          "model": "longcat_3p5b",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.2103",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "longcat_3p5b_cuda_p2.wav",
          "model": "longcat_3p5b",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3602",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "longcat_3p5b_cuda_p3.wav",
          "model": "longcat_3p5b",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4081",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "longcat_3p5b_cuda_p4.wav",
          "model": "longcat_3p5b",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3843",
          "wer": "0.1667",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "fish_s2",
      "name": "Fish Speech S2-Pro",
      "parameters_m": 4000.0,
      "approximate": false,
      "size_label": "4B",
      "size_note": "4B",
      "utmos": 4.310375,
      "wer": 0.08837500000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/fishaudio/s2-pro",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "fish_s2_cuda_p1.wav",
          "model": "fish_s2",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.1783",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "fish_s2_cuda_p2.wav",
          "model": "fish_s2",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.3286",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "fish_s2_cuda_p3.wav",
          "model": "fish_s2",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3987",
          "wer": "0.2424",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "fish_s2_cuda_p4.wav",
          "model": "fish_s2",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3359",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "higgs_v3",
      "name": "Higgs Audio v3 TTS",
      "parameters_m": 4000.0,
      "approximate": false,
      "size_label": "4B",
      "size_note": "4B",
      "utmos": 4.483025,
      "wer": 0.08080000000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/bosonai/higgs-audio-v3-tts-4b",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "higgs_v3_cuda_p1.wav",
          "model": "higgs_v3",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.4372",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "higgs_v3_cuda_p2.wav",
          "model": "higgs_v3",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4652",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "higgs_v3_cuda_p3.wav",
          "model": "higgs_v3",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5407",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "higgs_v3_cuda_p4.wav",
          "model": "higgs_v3",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4890",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "vibevoice_7b",
      "name": "VibeVoice 7B",
      "parameters_m": 7000.0,
      "approximate": false,
      "size_label": "7B",
      "size_note": "7B",
      "utmos": 3.495075,
      "wer": 0.083675,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/vibevoice/VibeVoice-7B",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "vibevoice_7b_cuda_p1.wav",
          "model": "vibevoice_7b",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "2.6372",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_7b_cuda_p2.wav",
          "model": "vibevoice_7b",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "3.7261",
          "wer": "0.1579",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_7b_cuda_p3.wav",
          "model": "vibevoice_7b",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.6232",
          "wer": "0.1212",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "vibevoice_7b_cuda_p4.wav",
          "model": "vibevoice_7b",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "3.9938",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "moss_tts",
      "name": "MOSS-TTS v1.0",
      "parameters_m": 8000.0,
      "approximate": false,
      "size_label": "8B",
      "size_note": "8B (Qwen3)",
      "utmos": 4.21845,
      "wer": 0.071775,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/OpenMOSS-Team/MOSS-TTS",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "moss_tts_cuda_p1.wav",
          "model": "moss_tts",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.0163",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_cuda_p2.wav",
          "model": "moss_tts",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.2720",
          "wer": "0.1053",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_cuda_p3.wav",
          "model": "moss_tts",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.4158",
          "wer": "0.1818",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_cuda_p4.wav",
          "model": "moss_tts",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.1697",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "moss_tts_v15",
      "name": "MOSS-TTS v1.5",
      "parameters_m": 8000.0,
      "approximate": false,
      "size_label": "8B",
      "size_note": "8B (Qwen3)",
      "utmos": 4.222075,
      "wer": 0.08080000000000001,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/OpenMOSS-Team/MOSS-TTS-v1.5",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "moss_tts_v15_cuda_p1.wav",
          "model": "moss_tts_v15",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "3.8375",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_v15_cuda_p2.wav",
          "model": "moss_tts_v15",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4165",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_v15_cuda_p3.wav",
          "model": "moss_tts_v15",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.3630",
          "wer": "0.2121",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "moss_tts_v15_cuda_p4.wav",
          "model": "moss_tts_v15",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.2713",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "miso",
      "name": "Miso TTS 8B",
      "parameters_m": 8200.0,
      "approximate": false,
      "size_label": "8.2B",
      "size_note": "8.2B",
      "utmos": 4.0718,
      "wer": 0.396575,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/MisoLabs/MisoTTS",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "miso_cuda_p1.wav",
          "model": "miso",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.1318",
          "wer": "0.2857",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "miso_cuda_p2.wav",
          "model": "miso",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4568",
          "wer": "0.3158",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "miso_cuda_p3.wav",
          "model": "miso",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "3.6810",
          "wer": "0.4848",
          "sim": "",
          "health": "gap"
        },
        {
          "dir": "linux-default",
          "wav": "miso_cuda_p4.wav",
          "model": "miso",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.0176",
          "wer": "0.5000",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "ming_tiny",
      "name": "Ming-3.0-TTS tiny",
      "parameters_m": 8980.0,
      "approximate": false,
      "size_label": "8.98B",
      "size_note": "8.98B (MoE, 1.3B active)",
      "utmos": 4.357225,
      "wer": 0.051775,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://huggingface.co/inclusionAI/Ming-3.0-tts-tiny",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "ming_tiny_cuda_p1.wav",
          "model": "ming_tiny",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3370",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ming_tiny_cuda_p2.wav",
          "model": "ming_tiny",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4003",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ming_tiny_cuda_p3.wav",
          "model": "ming_tiny",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.2952",
          "wer": "0.1515",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "ming_tiny_cuda_p4.wav",
          "model": "ming_tiny",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.3964",
          "wer": "0.0556",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    },
    {
      "id": "drama3",
      "name": "Fish Audio Drama 3 (ComfyUI port)",
      "parameters_m": 35900.0,
      "approximate": false,
      "size_label": "35.9B",
      "size_note": "35.9B (MoE, 3B active)",
      "utmos": 4.462350000000001,
      "wer": 0.0505,
      "wer_aggregation": "mean of four published prompt WERs; not corpus WER",
      "runtime": "Linux CUDA",
      "source": "https://github.com/PedroMarinhoDev/ComfyUI-Drama3",
      "local": false,
      "per_prompt": [
        {
          "dir": "linux-default",
          "wav": "drama3_cuda_p1.wav",
          "model": "drama3",
          "mode": "default",
          "prompt_id": "1",
          "utmos": "4.3715",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "drama3_cuda_p2.wav",
          "model": "drama3",
          "mode": "default",
          "prompt_id": "2",
          "utmos": "4.4675",
          "wer": "0.0000",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "drama3_cuda_p3.wav",
          "model": "drama3",
          "mode": "default",
          "prompt_id": "3",
          "utmos": "4.5239",
          "wer": "0.0909",
          "sim": "",
          "health": ""
        },
        {
          "dir": "linux-default",
          "wav": "drama3_cuda_p4.wav",
          "model": "drama3",
          "mode": "default",
          "prompt_id": "4",
          "utmos": "4.4865",
          "wer": "0.1111",
          "sim": "",
          "health": ""
        }
      ],
      "n": 4
    }
  ],
  "excluded": [
    {
      "name": "Flock Speech Goldfinch",
      "reason": "No audited result mapping"
    },
    {
      "name": "Flock Speech Junco",
      "reason": "No audited result mapping"
    },
    {
      "name": "KittenTTS Nano 0.1",
      "reason": "No numeric size"
    },
    {
      "name": "Paradee-8M v1.3",
      "reason": "No audited result mapping"
    },
    {
      "name": "Phonon 31m",
      "reason": "No audited result mapping"
    },
    {
      "name": "Phonon 40m",
      "reason": "No audited result mapping"
    },
    {
      "name": "PiVo-1.4M",
      "reason": "No audited result mapping"
    },
    {
      "name": "sanoTTS Heart-Nano",
      "reason": "No complete four-prompt run on one device"
    },
    {
      "name": "Voxtral 4B TTS",
      "reason": "No audited result mapping"
    },
    {
      "name": "Breeze TTS 2",
      "reason": "No audited result mapping"
    },
    {
      "name": "MioTTS 0.1B",
      "reason": "No audited result mapping"
    },
    {
      "name": "MioTTS 0.6B",
      "reason": "No audited result mapping"
    },
    {
      "name": "Qwen3-TTS 0.6B Base",
      "reason": "No audited result mapping"
    },
    {
      "name": "VoxMel 0.3B",
      "reason": "No audited result mapping"
    },
    {
      "name": "VoxMel 0.6B",
      "reason": "No audited result mapping"
    },
    {
      "name": "WavTTS",
      "reason": "No audited result mapping"
    },
    {
      "name": "Zonos2",
      "reason": "No audited result mapping"
    }
  ],
  "source_revision": "1dd3a81a50dd21fb2f4bce89df27eada996ba57f",
  "selection": "Same four English texts. Prefer the complete Linux CPU run; otherwise use the complete Linux CUDA run. Never combine devices within one model point. Default-voice mode only.",
  "raw_csv_sha256": "ff2dca6b6fce618144f0c86b5b54478c8e290c362dac220ec5b9142518e4447a",
  "note": "Four sentences per configuration. Published peer scores plus our earlier browser result; runtimes and scoring revisions differ.",
  "method": [
    "The same four English prompts are used for every point. CPU results are used when complete; otherwise the complete CUDA run is shown. No partial runs are combined.",
    "Peer scores come from tts-bench, not models rerun here. SmollerTTS is our earlier browser measurement. This overview uses a different test set from the 48-sentence comparison.",
    "Published WER is the mean of four prompt scores from the source recognizer, not the corpus WER in our 48-sentence test. The two sets of WER results should not be compared.",
    "Sizes are reported by each source; ~ means approximate. Components and counting conventions differ. Missing sizes or incomplete results are excluded.",
    "UTMOS predicts quality, not human preference. Four prompts are too few for a reliable ranking. The line only marks the best observed score up to each size."
  ],
  "sources": [
    {
      "name": "Scores",
      "url": "https://github.com/5uck1ess/tts-bench/blob/1dd3a81a50dd21fb2f4bce89df27eada996ba57f/scoring/scores.csv"
    },
    {
      "name": "Sizes",
      "url": "https://github.com/5uck1ess/tts-bench/blob/1dd3a81a50dd21fb2f4bce89df27eada996ba57f/README.md"
    },
    {
      "name": "Prompts",
      "url": "https://github.com/5uck1ess/tts-bench/blob/1dd3a81a50dd21fb2f4bce89df27eada996ba57f/scoring/prompts.py"
    }
  ]
}
