模型榜
本次来源更新未通过校验，保留上次数据。
把英语语音转为文字；错误率越低越好。
| 排名 | 模型 | 词错误率 |
|---|---|---|
| 1 | OtterAI/speech-v3.1闭源机构：OtterAI · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：6.42% Earnings22-Cleaned-AA-chunked：4.55% Gigaspeech-Cleaned：7.13% LS Clean：0.9% LS Other：1.89% SPGISpeech：1.56% URGENT24：8.07% Voice Arena Monsoon：3.02% Voxpopuli-AA-Cleaned：1.42% Private (scripted)：3.08% Private (conversational)：12.01% | 4.55% |
| 2 | zoom/scribe_v2_pro闭源机构：zoom · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：6.08% Earnings22-Cleaned-AA-chunked：4.12% Gigaspeech-Cleaned：7.75% LS Clean：1.11% LS Other：2.24% SPGISpeech：1.36% URGENT24：7.99% Voice Arena Monsoon：3% Voxpopuli-AA-Cleaned：3.08% Private (scripted)：2.92% Private (conversational)：12.1% | 4.7% |
| 3 | microsoft/azure-speech-07-2026闭源机构：microsoft · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：7.05% Earnings22-Cleaned-AA-chunked：5.2% Gigaspeech-Cleaned：7.48% LS Clean：1.23% LS Other：2.71% SPGISpeech：2.17% URGENT24：9.9% Voice Arena Monsoon：3% Voxpopuli-AA-Cleaned：1.65% Private (scripted)：2.12% Private (conversational)：12.71% | 5.02% |
| 4 | modulate/multilingual闭源机构：modulate · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：6.31% Earnings22-Cleaned-AA-chunked：6% Gigaspeech-Cleaned：6.91% LS Clean：0.87% LS Other：1.77% SPGISpeech：2.06% URGENT24：10.04% Voice Arena Monsoon：3.26% Voxpopuli-AA-Cleaned：3.55% Private (scripted)：2.56% Private (conversational)：12.3% | 5.06% |
| 5 | reson8/resonant-1闭源机构：reson8 · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：8.22% Earnings22-Cleaned-AA-chunked：5.48% Gigaspeech-Cleaned：7.38% LS Clean：0.92% LS Other：2.09% SPGISpeech：2.77% URGENT24：9.64% Voice Arena Monsoon：3.04% Voxpopuli-AA-Cleaned：2.19% Private (scripted)：2.96% Private (conversational)：12.61% | 5.21% |
| 6 | sophea/asr-k1 (preview)闭源机构：sophea · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：6.28% Earnings22-Cleaned-AA-chunked：5.56% Gigaspeech-Cleaned：7.96% LS Clean：0.9% LS Other：2.05% SPGISpeech：2.11% URGENT24：10.07% Voice Arena Monsoon：3.25% Voxpopuli-AA-Cleaned：3.35% Private (scripted)：2.86% Private (conversational)：13.16% | 5.23% |
| 7 | elevenlabs/scribe_v2闭源机构：elevenlabs · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：8.19% Earnings22-Cleaned-AA-chunked：4.8% Gigaspeech-Cleaned：7.63% LS Clean：1.06% LS Other：2.33% SPGISpeech：3.01% URGENT24：9.2% Voice Arena Monsoon：3.32% Voxpopuli-AA-Cleaned：1.41% Private (scripted)：3% Private (conversational)：13.66% | 5.24% |
| 8 | assemblyai/universal-3-5-pro闭源机构：assemblyai · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：10.63% Earnings22-Cleaned-AA-chunked：6.05% Gigaspeech-Cleaned：7.62% LS Clean：1.1% LS Other：2.11% SPGISpeech：1.81% URGENT24：9.08% Voice Arena Monsoon：3.49% Voxpopuli-AA-Cleaned：1.92% Private (scripted)：2.66% Private (conversational)：12.8% | 5.39% |
| 9 | Qwen/Qwen3-ASR-1.7B-hf开源机构：Qwen · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：8.31% Earnings22-Cleaned-AA-chunked：5.84% Gigaspeech-Cleaned：7.22% LS Clean：1.26% LS Other：2.94% SPGISpeech：2.57% URGENT24：9.93% Voice Arena Monsoon：3.5% Voxpopuli-AA-Cleaned：2.85% Private (scripted)：2.87% Private (conversational)：12.14% | 5.4% |
| 10 | ibm-granite/granite-speech-5.0-2b开源机构：ibm-granite · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：6.96% Earnings22-Cleaned-AA-chunked：6.22% Gigaspeech-Cleaned：8.14% LS Clean：0.99% LS Other：2.21% SPGISpeech：3.36% URGENT24：10.22% Voice Arena Monsoon：3.94% Voxpopuli-AA-Cleaned：3.52% Private (scripted)：3.16% Private (conversational)：13.79% | 5.68% |
| 11 | smallestai/pulse闭源机构：smallestai · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：9.11% Earnings22-Cleaned-AA-chunked：4.74% Gigaspeech-Cleaned：8.18% LS Clean：1.3% LS Other：2.7% SPGISpeech：1.91% URGENT24：10.44% Voice Arena Monsoon：3.52% Voxpopuli-AA-Cleaned：3.85% Private (scripted)：3.58% Private (conversational)：13.5% | 5.71% |
| 12 | CohereLabs/cohere-transcribe-03-2026开源机构：CohereLabs · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.02% Earnings22-Cleaned-AA-chunked：7.88% Gigaspeech-Cleaned：7.88% LS Clean：0.97% LS Other：2.05% SPGISpeech：2.74% URGENT24：9.5% Voice Arena Monsoon：3.44% Voxpopuli-AA-Cleaned：5.38% Private (scripted)：3.2% Private (conversational)：13.46% | 5.77% |
| 13 | sprag/symphony闭源机构：sprag · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：9.02% Earnings22-Cleaned-AA-chunked：6.01% Gigaspeech-Cleaned：7.52% LS Clean：1% LS Other：2.19% SPGISpeech：2.21% URGENT24：12.64% Voice Arena Monsoon：3.55% Voxpopuli-AA-Cleaned：3.33% Private (scripted)：3.99% Private (conversational)：12.78% | 5.84% |
| 14 | HojoAI/Hojo-ASR-V1开源机构：HojoAI · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.54% Earnings22-Cleaned-AA-chunked：6.97% Gigaspeech-Cleaned：6.14% LS Clean：1.3% LS Other：3.24% SPGISpeech：1.74% URGENT24：12.33% Voice Arena Monsoon：4.63% Voxpopuli-AA-Cleaned：3.11% Private (scripted)：3.99% Private (conversational)：13.4% | 5.85% |
| 15 | TheStageAI/thewhisper-large-v3-turbo开源机构：TheStageAI · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.32% Earnings22-Cleaned-AA-chunked：7.18% Gigaspeech-Cleaned：7.84% LS Clean：1.41% LS Other：3.2% SPGISpeech：1.62% URGENT24：10.82% Voice Arena Monsoon：3.72% Voxpopuli-AA-Cleaned：4.02% Private (scripted)：3.93% Private (conversational)：13.98% | 5.91% |
| 16 | bosonai/higgs-audio-v3-stt开源机构：bosonai · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.19% Earnings22-Cleaned-AA-chunked：7.04% Gigaspeech-Cleaned：7.45% LS Clean：1.08% LS Other：2.61% SPGISpeech：2.07% URGENT24：11.88% Voice Arena Monsoon：3.97% Voxpopuli-AA-Cleaned：3.73% Private (scripted)：3.74% Private (conversational)：14.34% | 5.92% |
| 17 | Edge0/ARK-ASR-3B开源机构：Edge0 · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.89% Earnings22-Cleaned-AA-chunked：8.07% Gigaspeech-Cleaned：6.94% LS Clean：1.04% LS Other：2.37% SPGISpeech：2.46% URGENT24：13.29% Voice Arena Monsoon：3.48% Voxpopuli-AA-Cleaned：3.55% Private (scripted)：3.11% Private (conversational)：12.9% | 5.92% |
| 18 | Edge0/ARK-ASR-0.6B开源机构：Edge0 · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：8.61% Earnings22-Cleaned-AA-chunked：6.75% Gigaspeech-Cleaned：7.37% LS Clean：1.48% LS Other：3.44% SPGISpeech：2.63% URGENT24：12.49% Voice Arena Monsoon：3.43% Voxpopuli-AA-Cleaned：2.76% Private (scripted)：3.68% Private (conversational)：12.57% | 5.93% |
| 19 | nvidia/canary-qwen-2.5b开源机构：nvidia · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.91% Earnings22-Cleaned-AA-chunked：6.41% Gigaspeech-Cleaned：7.8% LS Clean：1.23% LS Other：2.63% SPGISpeech：1.7% URGENT24：13.02% Voice Arena Monsoon：3.6% Voxpopuli-AA-Cleaned：4.14% Private (scripted)：3.47% Private (conversational)：13.41% | 5.94% |
| 20 | ibm-granite/granite-speech-4.1-2b开源机构：ibm-granite · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.06% Earnings22-Cleaned-AA-chunked：6.69% Gigaspeech-Cleaned：8.18% LS Clean：1.03% LS Other：2.19% SPGISpeech：3.47% URGENT24：11.25% Voice Arena Monsoon：4.12% Voxpopuli-AA-Cleaned：4.18% Private (scripted)：3.16% Private (conversational)：14.19% | 5.96% |
| 21 | openai/gpt-transcribe闭源机构：openai · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：8.95% Earnings22-Cleaned-AA-chunked：6.61% Gigaspeech-Cleaned：7.76% LS Clean：1.31% LS Other：3.39% SPGISpeech：2.23% URGENT24：12.86% Voice Arena Monsoon：3.63% Voxpopuli-AA-Cleaned：2.65% Private (scripted)：3.24% Private (conversational)：12.9% | 5.96% |
| 22 | nvidia/parakeet-tdt-0.6b-v2 (fast-gpu-asr)开源机构：nvidia · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：9.03% Earnings22-Cleaned-AA-chunked：6.84% Gigaspeech-Cleaned：8.16% LS Clean：1.27% LS Other：2.62% SPGISpeech：1.87% URGENT24：11.64% Voice Arena Monsoon：3.72% Voxpopuli-AA-Cleaned：3.85% Private (scripted)：3.48% Private (conversational)：13.33% | 5.98% |
| 23 | gladia/solaria-3闭源机构：gladia · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：9.89% Earnings22-Cleaned-AA-chunked：5.94% Gigaspeech-Cleaned：7.32% LS Clean：1.31% LS Other：2.99% SPGISpeech：2.68% URGENT24：13.73% Voice Arena Monsoon：3.7% Voxpopuli-AA-Cleaned：2.76% Private (scripted)：3.12% Private (conversational)：12.44% | 5.99% |
| 24 | ibm-granite/granite-speech-5.0-470m-turboctc-nc受限开放机构：ibm-granite · 模型原始链接 模型许可证：cc-by-nc-sa-4.0 来源模型许可证：cc-by-nc-sa-4.0，含非商用限制，不等于可自由商用。 AMI-Cleaned：7.31% Earnings22-Cleaned-AA-chunked：8.23% Gigaspeech-Cleaned：8.37% LS Clean：1.45% LS Other：2.67% SPGISpeech：1.75% URGENT24：10.22% Voice Arena Monsoon：4.3% Voxpopuli-AA-Cleaned：4.67% Private (scripted)：3.81% Private (conversational)：13.49% | 6.02% |
| 25 | nvidia/parakeet-tdt-0.6b-v2开源机构：nvidia · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：9.09% Earnings22-Cleaned-AA-chunked：6.76% Gigaspeech-Cleaned：8.15% LS Clean：1.28% LS Other：2.73% SPGISpeech：1.94% URGENT24：11.42% Voice Arena Monsoon：3.89% Voxpopuli-AA-Cleaned：3.78% Private (scripted)：3.62% Private (conversational)：13.58% | 6.02% |
| 26 | nvidia/parakeet-tdt-0.6b-v3 (fast-gpu-asr)开源机构：nvidia · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：9.01% Earnings22-Cleaned-AA-chunked：5.81% Gigaspeech-Cleaned：7.75% LS Clean：1.45% LS Other：3.08% SPGISpeech：3.56% URGENT24：11.5% Voice Arena Monsoon：3.96% Voxpopuli-AA-Cleaned：3.04% Private (scripted)：4.54% Private (conversational)：13.17% | 6.08% |
| 27 | ibm-granite/granite-speech-4.1-2b-nar开源机构：ibm-granite · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：6.96% Earnings22-Cleaned-AA-chunked：6.67% Gigaspeech-Cleaned：8.61% LS Clean：1.04% LS Other：2.4% SPGISpeech：3.23% URGENT24：11.19% Voice Arena Monsoon：4.24% Voxpopuli-AA-Cleaned：4.25% Private (scripted)：4.21% Private (conversational)：14.34% | 6.1% |
| 28 | OpenMOSS-Team/MOSS-Transcribe-Diarize开源机构：OpenMOSS-Team · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：8.33% Earnings22-Cleaned-AA-chunked：6.09% Gigaspeech-Cleaned：7.68% LS Clean：1.66% LS Other：4.16% SPGISpeech：1.7% URGENT24：12.74% Voice Arena Monsoon：3.87% Voxpopuli-AA-Cleaned：3.6% Private (scripted)：4.68% Private (conversational)：13.4% | 6.17% |
| 29 | nvidia/parakeet-tdt-0.6b-v3开源机构：nvidia · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：9.42% Earnings22-Cleaned-AA-chunked：5.85% Gigaspeech-Cleaned：7.99% LS Clean：1.52% LS Other：3.13% SPGISpeech：3.63% URGENT24：11.52% Voice Arena Monsoon：4.14% Voxpopuli-AA-Cleaned：3.05% Private (scripted)：4.27% Private (conversational)：13.44% | 6.18% |
| 30 | Qwen/Qwen3-ASR-0.6B-hf开源机构：Qwen · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：9.33% Earnings22-Cleaned-AA-chunked：7.83% Gigaspeech-Cleaned：7.62% LS Clean：1.7% LS Other：4.01% SPGISpeech：2.74% URGENT24：11.57% Voice Arena Monsoon：4.23% Voxpopuli-AA-Cleaned：2.9% Private (scripted)：3.92% Private (conversational)：12.74% | 6.24% |
| 31 | meta/muse-voice-transcribe闭源机构：meta · 模型原始链接 模型许可证：Proprietary 来源将此模型标记为专有许可。 AMI-Cleaned：9.2% Earnings22-Cleaned-AA-chunked：6.01% Gigaspeech-Cleaned：7.75% LS Clean：1.84% LS Other：4.51% SPGISpeech：3.21% URGENT24：13.03% Voice Arena Monsoon：4.16% Voxpopuli-AA-Cleaned：2.02% Private (scripted)：4.52% Private (conversational)：13.04% | 6.3% |
| 32 | ibm-granite/granite-speech-5.0-470m-turboctc开源机构：ibm-granite · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.71% Earnings22-Cleaned-AA-chunked：6.75% Gigaspeech-Cleaned：8.74% LS Clean：1.48% LS Other：2.71% SPGISpeech：3.22% URGENT24：11.79% Voice Arena Monsoon：4.75% Voxpopuli-AA-Cleaned：4.9% Private (scripted)：4.13% Private (conversational)：14.03% | 6.38% |
| 33 | mistralai/Voxtral-Small-24B-2507开源机构：mistralai · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：13.19% Earnings22-Cleaned-AA-chunked：7.07% Gigaspeech-Cleaned：8.33% LS Clean：1.23% LS Other：2.8% SPGISpeech：1.84% URGENT24：12.35% Voice Arena Monsoon：3.49% Voxpopuli-AA-Cleaned：2% Private (scripted)：3.22% Private (conversational)：14.7% | 6.38% |
| 34 | ibm-granite/granite-4.0-1b-speech开源机构：ibm-granite · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：7.38% Earnings22-Cleaned-AA-chunked：8.53% Gigaspeech-Cleaned：8.44% LS Clean：1.11% LS Other：2.5% SPGISpeech：3.55% URGENT24：11.99% Voice Arena Monsoon：4.84% Voxpopuli-AA-Cleaned：4.41% Private (scripted)：3.56% Private (conversational)：14.22% | 6.41% |
| 35 | nvidia/nemotron-speech-streaming-en-0.6b待核实机构：nvidia · 模型原始链接 模型许可证：nvidia-open-model-license 来源模型许可证：nvidia-open-model-license。自定义或不完整许可尚未核实，不自动归为开源。 AMI-Cleaned：8.84% Earnings22-Cleaned-AA-chunked：9.45% Gigaspeech-Cleaned：7.9% LS Clean：1.89% LS Other：4.35% SPGISpeech：2.69% URGENT24：12.54% Voice Arena Monsoon：4.44% Voxpopuli-AA-Cleaned：2.4% Private (scripted)：4.79% Private (conversational)：13.28% | 6.6% |
| 36 | distil-whisper/distil-large-v3.5开源机构：distil-whisper · 模型原始链接 模型许可证：mit 来源模型许可证：mit。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：12.09% Earnings22-Cleaned-AA-chunked：7.7% Gigaspeech-Cleaned：8.21% LS Clean：1.94% LS Other：4.52% SPGISpeech：2.62% URGENT24：12.42% Voice Arena Monsoon：3.6% Voxpopuli-AA-Cleaned：2.52% Private (scripted)：4.34% Private (conversational)：14.43% | 6.76% |
| 37 | soundsgoodai/Zipformer-cr-ctc-transducer-XL-290M (fast-gpu-asr, transducer_modified_beam_search)受限开放机构：soundsgoodai · 模型原始链接 模型许可证：cc-by-nc-4.0 来源模型许可证：cc-by-nc-4.0，含非商用限制，不等于可自由商用。 AMI-Cleaned：10.23% Earnings22-Cleaned-AA-chunked：7.9% Gigaspeech-Cleaned：8.3% LS Clean：1.31% LS Other：3.03% SPGISpeech：1.64% URGENT24：13.67% Voice Arena Monsoon：5.57% Voxpopuli-AA-Cleaned：4.32% Private (scripted)：4.32% Private (conversational)：14.74% | 6.82% |
| 38 | FermionResearch/Phonon-2开源机构：FermionResearch · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：9.32% Earnings22-Cleaned-AA-chunked：6.96% Gigaspeech-Cleaned：8.35% LS Clean：1.73% LS Other：3.92% SPGISpeech：3.7% URGENT24：14.9% Voice Arena Monsoon：4.91% Voxpopuli-AA-Cleaned：2.46% Private (scripted)：5.23% Private (conversational)：14.14% | 6.87% |
| 39 | openai/whisper-large-v3开源机构：openai · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：13.63% Earnings22-Cleaned-AA-chunked：7.99% Gigaspeech-Cleaned：8.35% LS Clean：1.56% LS Other：3.52% SPGISpeech：2.71% URGENT24：13.73% Voice Arena Monsoon：3.95% Voxpopuli-AA-Cleaned：4.53% Private (scripted)：3.61% Private (conversational)：15.13% | 7.16% |
| 40 | FunAudioLLM/Fun-ASR-Nano-2512-hf开源机构：FunAudioLLM · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：10.35% Earnings22-Cleaned-AA-chunked：10.87% Gigaspeech-Cleaned：8.41% LS Clean：1.47% LS Other：4.07% SPGISpeech：2.05% URGENT24：13.93% Voice Arena Monsoon：5.32% Voxpopuli-AA-Cleaned：3.4% Private (scripted)：4.95% Private (conversational)：14.83% | 7.24% |
| 41 | mistralai/Voxtral-Mini-3B-2507开源机构：mistralai · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：13.57% Earnings22-Cleaned-AA-chunked：7.66% Gigaspeech-Cleaned：8.7% LS Clean：1.48% LS Other：3.63% SPGISpeech：2.15% URGENT24：15.66% Voice Arena Monsoon：4.9% Voxpopuli-AA-Cleaned：2.22% Private (scripted)：3.73% Private (conversational)：16% | 7.25% |
| 42 | nvidia/canary-1b-v2开源机构：nvidia · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：13.03% Earnings22-Cleaned-AA-chunked：7.28% Gigaspeech-Cleaned：9.12% LS Clean：1.76% LS Other：3.07% SPGISpeech：2.03% URGENT24：14.56% Voice Arena Monsoon：4.97% Voxpopuli-AA-Cleaned：4.39% Private (scripted)：4.65% Private (conversational)：15.16% | 7.27% |
| 43 | nvidia/parakeet-ctc-1.1b开源机构：nvidia · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：12.23% Earnings22-Cleaned-AA-chunked：9.66% Gigaspeech-Cleaned：8.74% LS Clean：1.52% LS Other：3.23% SPGISpeech：3.53% URGENT24：14.88% Voice Arena Monsoon：5.23% Voxpopuli-AA-Cleaned：3.24% Private (scripted)：4.44% Private (conversational)：14.7% | 7.4% |
| 44 | openai/whisper-large-v3-turbo开源机构：openai · 模型原始链接 模型许可证：mit 来源模型许可证：mit。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：13.88% Earnings22-Cleaned-AA-chunked：8.09% Gigaspeech-Cleaned：8.47% LS Clean：2.13% LS Other：3.71% SPGISpeech：2.79% URGENT24：13.69% Voice Arena Monsoon：4.77% Voxpopuli-AA-Cleaned：7.02% Private (scripted)：4.05% Private (conversational)：15.4% | 7.64% |
| 45 | espnet/owsm_ctc_v4_1B开源机构：espnet · 模型原始链接 模型许可证：cc-by-4.0 来源模型许可证：cc-by-4.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：11.1% Earnings22-Cleaned-AA-chunked：10.31% Gigaspeech-Cleaned：9.29% LS Clean：2.11% LS Other：4.38% SPGISpeech：2.31% URGENT24：16.12% Voice Arena Monsoon：6.66% Voxpopuli-AA-Cleaned：4.09% Private (scripted)：6.41% Private (conversational)：16.1% | 8.08% |
| 46 | facebook/omniASR-LLM-7B-v2开源机构：facebook · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：13.86% Earnings22-Cleaned-AA-chunked：11.44% Gigaspeech-Cleaned：9.2% LS Clean：1.35% LS Other：3.25% SPGISpeech：3.22% URGENT24：14.84% Voice Arena Monsoon：4.87% Voxpopuli-AA-Cleaned：4.03% Private (scripted)：7.56% Private (conversational)：15.85% | 8.13% |
| 47 | microsoft/VibeVoice-ASR-HF开源机构：microsoft · 模型原始链接 模型许可证：mit 来源模型许可证：mit。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：12.02% Earnings22-Cleaned-AA-chunked：7.97% Gigaspeech-Cleaned：7.87% LS Clean：1.56% LS Other：4.95% SPGISpeech：3.47% URGENT24：19.5% Voice Arena Monsoon：4.74% Voxpopuli-AA-Cleaned：2.02% Private (scripted)：11.09% Private (conversational)：14.62% | 8.16% |
| 48 | Edge0/Audio8-ASR-0.1B开源机构：Edge0 · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：11% Earnings22-Cleaned-AA-chunked：12.53% Gigaspeech-Cleaned：8.48% LS Clean：2.71% LS Other：6.6% SPGISpeech：3.73% URGENT24：14.97% Voice Arena Monsoon：6.18% Voxpopuli-AA-Cleaned：4.39% Private (scripted)：5.85% Private (conversational)：14.79% | 8.29% |
| 49 | soundsgoodai/Zipformer-cr-ctc-transducer-XL-290M (fast-gpu-asr, ctc_greedy_search)受限开放机构：soundsgoodai · 模型原始链接 模型许可证：cc-by-nc-4.0 来源模型许可证：cc-by-nc-4.0，含非商用限制，不等于可自由商用。 AMI-Cleaned：11.63% Earnings22-Cleaned-AA-chunked：9.24% Gigaspeech-Cleaned：9.52% LS Clean：1.71% LS Other：3.9% SPGISpeech：2.75% URGENT24：17.67% Voice Arena Monsoon：7.76% Voxpopuli-AA-Cleaned：4.53% Private (scripted)：6.99% Private (conversational)：16.47% | 8.38% |
| 50 | mistralai/Voxtral-Mini-4B-Realtime-2602开源机构：mistralai · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：13.34% Earnings22-Cleaned-AA-chunked：9.31% Gigaspeech-Cleaned：8.8% LS Clean：1.62% LS Other：4.94% SPGISpeech：2.23% URGENT24：17.95% Voice Arena Monsoon：8.86% Voxpopuli-AA-Cleaned：2.6% Private (scripted)：5.76% Private (conversational)：17.45% | 8.44% |
| 51 | abr-ai/niagara-84m-batch.en待核实机构：abr-ai · 模型原始链接 模型许可证：applied-brain-research-open-license 来源模型许可证：applied-brain-research-open-license。自定义或不完整许可尚未核实，不自动归为开源。 AMI-Cleaned：12.01% Earnings22-Cleaned-AA-chunked：10.83% Gigaspeech-Cleaned：10.41% LS Clean：3.09% LS Other：7.6% SPGISpeech：2.61% URGENT24：20.16% Voice Arena Monsoon：8.82% Voxpopuli-AA-Cleaned：4.33% Private (scripted)：8.7% Private (conversational)：18.08% | 9.69% |
| 52 | nvidia/nemotron-3.5-asr-streaming-0.6b待核实机构：nvidia · 模型原始链接 模型许可证：OpenMDW-1.1 来源模型许可证：OpenMDW-1.1。自定义或不完整许可尚未核实，不自动归为开源。 AMI-Cleaned：13.43% Earnings22-Cleaned-AA-chunked：15.3% Gigaspeech-Cleaned：9.86% LS Clean：2.83% LS Other：6.79% SPGISpeech：3.27% URGENT24：20.14% Voice Arena Monsoon：7.3% Voxpopuli-AA-Cleaned：4.24% Private (scripted)：7.84% Private (conversational)：16.3% | 9.75% |
| 53 | abr-ai/niagara-38m-batch.en待核实机构：abr-ai · 模型原始链接 模型许可证：applied-brain-research-open-license 来源模型许可证：applied-brain-research-open-license。自定义或不完整许可尚未核实，不自动归为开源。 AMI-Cleaned：13.87% Earnings22-Cleaned-AA-chunked：12.94% Gigaspeech-Cleaned：11.38% LS Clean：3.81% LS Other：9.28% SPGISpeech：3.08% URGENT24：22.88% Voice Arena Monsoon：10.41% Voxpopuli-AA-Cleaned：5.16% Private (scripted)：10.05% Private (conversational)：19.65% | 11.14% |
| 54 | facebook/omniASR-CTC-7B-v2开源机构：facebook · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：17.44% Earnings22-Cleaned-AA-chunked：18.3% Gigaspeech-Cleaned：12.95% LS Clean：2.12% LS Other：5.05% SPGISpeech：5.16% URGENT24：18.68% Voice Arena Monsoon：7.08% Voxpopuli-AA-Cleaned：4.43% Private (scripted)：14.66% Private (conversational)：20.62% | 11.5% |
| 55 | google/gemma-4-E4B-it开源机构：google · 模型原始链接 模型许可证：apache-2.0 来源模型许可证：apache-2.0。按公开权重口径展示，使用前请核对原始许可。 AMI-Cleaned：20.57% Earnings22-Cleaned-AA-chunked：14.16% Gigaspeech-Cleaned：11.74% LS Clean：3.85% LS Other：9.97% SPGISpeech：5.49% URGENT24：26.38% Voice Arena Monsoon：6.19% Voxpopuli-AA-Cleaned：4.09% Private (scripted)：8.88% Private (conversational)：19.17% | 11.86% |
| 56 | abr-ai/niagara-19m-batch.en待核实机构：abr-ai · 模型原始链接 模型许可证：applied-brain-research-open-license 来源模型许可证：applied-brain-research-open-license。自定义或不完整许可尚未核实，不自动归为开源。 AMI-Cleaned：16.17% Earnings22-Cleaned-AA-chunked：16.1% Gigaspeech-Cleaned：12.72% LS Clean：4.74% LS Other：11.02% SPGISpeech：3.67% URGENT24：25.4% Voice Arena Monsoon：13.15% Voxpopuli-AA-Cleaned：5.73% Private (scripted)：11.87% Private (conversational)：22.1% | 12.97% |
| 57 | abr-ai/niagara-9m-batch.en待核实机构：abr-ai · 模型原始链接 模型许可证：applied-brain-research-open-license 来源模型许可证：applied-brain-research-open-license。自定义或不完整许可尚未核实，不自动归为开源。 AMI-Cleaned：20.94% Earnings22-Cleaned-AA-chunked：21.46% Gigaspeech-Cleaned：14.97% LS Clean：6.45% LS Other：14.53% SPGISpeech：5.16% URGENT24：29.47% Voice Arena Monsoon：17.11% Voxpopuli-AA-Cleaned：7.33% Private (scripted)：15.15% Private (conversational)：25.62% | 16.2% |
评测口径与数据来源
保留来源默认短音频英语榜的名次与 Average WER，包括来源已公开的私有测试集汇总分。不是中文识别榜；没有读取私有音频、转录或测试题。