Leaderboard
WER
werRecord
- Category
- speech-asr
- Scored models
- 688
Scores
| Model | Score |
|---|---|
| — | 100 |
| facebook/mms-300m | 100 |
| facebook/mms-300m | 100 |
| facebook/mms-300m | 100 |
| sulaimank/w2vbert-luganda-waxal-stage2 | 100 |
| facebook/mms-300m | 100 |
| — | 100 |
| facebook/mms-300m | 100 |
| facebook/mms-300m | 100 |
| openai/whisper-small | 99.73 |
| microsoft/Phi-4-multimodal-instruct | 99.63 |
| microsoft/Phi-4-multimodal-instruct | 99.63 |
| openai/whisper-small | 98 |
| openai/whisper-small | 95.92 |
| openai/whisper-small | 95.26 |
| — | 95.24 |
| openai/whisper-small | 94.94 |
| openai/whisper-small | 93.83 |
| UsefulSensors/moonshine-tiny-vi | 91.89 |
| UsefulSensors/moonshine-tiny-vi | 91.89 |
| openai/whisper-tiny | 90.98 |
| openai/whisper-tiny | 87.12 |
| openai/whisper-tiny | 87.12 |
| openai/whisper-medium | 86.99 |
Remaining fields
Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.
- category
- speech-asr
- id
- wer
- name
- WER
- schema version
- local-ai-registry/v1
rows
| variant | conf | context | model id | org | rank | root | score |
|---|---|---|---|---|---|---|---|
| wav2vec2-ljspeech-gruut | low | Wer | Unknown | bookbot | 1 | — | 100 |
| mms-300m-waxal-lug | low | Wer | Unknown | waxal-benchmarking | 2 | facebook/mms-300m | 100 |
| mms-300m-waxal-sna | low | Wer | Unknown | waxal-benchmarking | 3 | facebook/mms-300m | 100 |
| mms-300m-waxal-lin | low | Wer | Unknown | waxal-benchmarking | 4 | facebook/mms-300m | 100 |
| w2vbert-luganda-waxal-punct | low | Wer Keep | Unknown | sulaimank | 5 | sulaimank/w2vbert-luganda-waxal-stage2 | 100 |
| mms-300m-waxal-ach | low | Wer | Unknown | waxal-benchmarking | 6 | facebook/mms-300m | 100 |
| phoneme-scorer-v2-wav2vec2 | low | Wer | Unknown | ct-vikramanantha | 7 | — | 100 |
| mms-300m-waxal-nyn | low | Wer | Unknown | waxal-benchmarking | 8 | facebook/mms-300m | 100 |
| mms-300m-waxal-mas | low | Wer | Unknown | waxal-benchmarking | 9 | facebook/mms-300m | 100 |
| whisper-small-lingala-qlora | medium | Wer | whisper-small | Elod1e | 10 | openai/whisper-small | 99.73 |
| Phi-4-multimodal-instruct | low | fleurs-ko2en-cot | fleurs-en2ko-cot | phi-4-multimodal-instruct | microsoft | 11 | microsoft/Phi-4-multimodal-instruct | 99.63 |
| Phi-4-multimodal-instruct | low | fleurs-ko2en-cot | fleurs-en2ko-cot | phi-4-multimodal-instruct | Lexius | 12 | microsoft/Phi-4-multimodal-instruct | 99.63 |
| whisper-small-finetuned-iwslt-irish-welsh-curriculum-synthetic-b | low | Wer | whisper-small | jjnhuang | 13 | openai/whisper-small | 98 |
| whisper-small-finetuned-iwslt-irish-welsh-50-50-synthetic | medium | Wer | whisper-small | jjnhuang | 14 | openai/whisper-small | 95.92 |
| whisper-small-finetuned-iwslt-irish-transfer-welsh-synthetic | medium | Wer | whisper-small | jjnhuang | 15 | openai/whisper-small | 95.26 |
| wav2vec2-large-xlsr-53-th | low | WER PyThaiNLP 2.3.1 | Unknown | airesearch | 16 | — | 95.24 |
| whisper-small-finetuned-iwslt-irish-synthetic | medium | Wer | whisper-small | jjnhuang | 17 | openai/whisper-small | 94.94 |
| whisper-small-finetuned-iwslt-irish-welsh-curriculum-synthetic | medium | Wer | whisper-small | jjnhuang | 18 | openai/whisper-small | 93.83 |
| moonshine-tiny-vi | low | Fleurs (WER) ↓ | moonshine-tiny-vi | UsefulSensors | 19 | UsefulSensors/moonshine-tiny-vi | 91.89 |
| moonshine-tiny-vi-ONNX | low | Fleurs (WER) ↓ | moonshine-tiny-vi | onnx-community | 20 | UsefulSensors/moonshine-tiny-vi | 91.89 |
| whisper-tiny-ja | medium | Wer | whisper-tiny | zubokol | 21 | openai/whisper-tiny | 90.98 |
| cali-whisper-tiny.en-drop-004-production | low | Wer Ortho | whisper-tiny | 1Developer | 22 | openai/whisper-tiny | 87.12 |
| cali-whisper-tiny.en-drop-003-production | low | Wer Ortho | whisper-tiny | 1Developer | 23 | openai/whisper-tiny | 87.12 |
| whisper-medium-hindi-hinglish-asr-fine-tuned | medium | Relative WER Improvement | whisper-medium | shujaAK | 24 | openai/whisper-medium | 86.99 |
| whisper-medium-hindi-hinglish-asr-fine-tuned | medium | WER Improvement | whisper-medium | shujaAK | 25 | openai/whisper-medium | 86.99 |
| ATOM | low | WER (Armenian) | Unknown | Chillarmo | 26 | — | 86.6 |
| wav2vec2-large-xlsr-tamil-commonvoice | medium | Wer | Unknown | nikhil6041 | 27 | — | 85.12 |
| whisper-small-yemeni | low | WER (%) | whisper-small | tahaalselwii | 28 | openai/whisper-small | 84.32 |
| wav2vec2-xlsr-multilingual-56 | high | WER | Unknown | voidful | 29 | — | 82.5 |
| wav2vec2-large-xlsr-53-chinese-zh-cn | high | WER | Unknown | jonatasgrosman | 30 | jonatasgrosman/wav2vec2-large-xlsr-53-chinese-zh-cn | 82.37 |
| parakeet-tdt-0.6b-v3 | low | WER (LibriSpeech) | parakeet-tdt-0-6b-v3 | sonic-speech | 31 | nvidia/parakeet-tdt-0.6b-v3 | 82 |
| parakeet-tdt-0.6b-v3-int8 | medium | WER (LibriSpeech) | parakeet-tdt-0-6b-v3 | sonic-speech | 32 | sonic-speech/parakeet-tdt-0.6b-v3 | 82 |
| wav2vec2-large-xlsr-53-japanese | high | WER | Unknown | jonatasgrosman | 33 | jonatasgrosman/wav2vec2-large-xlsr-53-japanese | 81.8 |
| nemotron-3.5-nb-NO | medium | WER (offline, full-context, normalized) on `LokaalHub/nb-NO-asr-cv` test | nemotron-3-5-asr-streaming-0-6b | LokaalHub | 34 | nvidia/nemotron-3.5-asr-streaming-0.6b | 76.09 |
| qwen3-asr-0.6b-arabic-dialectal | high | Zero-shot WER | qwen3-asr-0-6b-hf | oddadmix | 35 | Qwen/Qwen3-ASR-0.6B-hf | 75.6 |
| wav2vec2-base-german | medium | Wer | Unknown | aware-ai | 36 | facebook/wav2vec2-base | 74.28 |
| wav2vec2-large-japanese | high | WER | Unknown | NTQAI | 37 | — | 73.1 |
| wav2vec2-large-chinese-zh-cn | high | WER | Unknown | wbbbbb | 38 | — | 70.47 |
| wav2vec2-large-xls-r-300m-armenian | medium | Wer | Unknown | infinitejoy | 39 | — | 69.42 |
| whisper-large-v3-turbo-singlish | low | WER | whisper-large-v3-turbo | mjwong | 40 | openai/whisper-large-v3-turbo | 68.79 |
| wav2vec2-large-mms-1b-yoruba-test | medium | Wer | mms-1b-all | ogbi | 41 | facebook/mms-1b-all | 68.02 |
| whisper-tiny-yoruba1 | medium | Wer | Unknown | EYEDOL | 42 | EYEDOL/whisper-tiny-yoruba | 67.55 |
| whisper-small-burmese-v2 | medium | Wer | whisper-small | myatsu | 43 | myatsu/whisper-small-burmese | 67.36 |
| whisper-tiny-grpo-5e2f49e-cv22-tiny-sftadaptive | medium | WER | whisper-tiny | wrice | 44 | openai/whisper-tiny | 67.1 |
| moulsot.v0.3 | medium | WER | qwen3-asr-1-7b | atlasia | 45 | Qwen/Qwen3-ASR-1.7B | 66.78 |
| moonshine-base-bn | low | WER (%) | moonshine-base | Sanjidh090 | 46 | UsefulSensors/moonshine-base | 66.71 |
| whisper-tiny-yoruba2 | medium | Wer | Unknown | EYEDOL | 47 | EYEDOL/whisper-tiny-yoruba1 | 66.05 |
| moonshine-tiny-ar-ONNX | low | Fleurs (WER) ↓ | moonshine-tiny-ar | onnx-community | 48 | UsefulSensors/moonshine-tiny-ar | 66.01 |
| moonshine-tiny-ar | low | Fleurs (WER) ↓ | moonshine-tiny-ar | UsefulSensors | 49 | UsefulSensors/moonshine-tiny-ar | 66.01 |
| whisper-large-v3_finetuned_ugandan_luganda_nonstandard_speech_v1.0 | low | Standard WER | Unknown | cdli | 50 | — | 66 |
| cdli-qwen3-asr-lg-atypical-stage3-1p7b-base | medium | WER | qwen3-asr-1-7b | KasuleTrevor | 51 | KasuleTrevor/cdli-qwen3-asr-lg-typical-1p7b-base-finetune | 65.91 |
| whisper-tiny-grpo-4bddbd7-cv22-tiny-klbeta02 | medium | WER | whisper-tiny | wrice | 52 | openai/whisper-tiny | 65.6 |
| wav2vec2-large-xls-r-300m-assamese | medium | Wer | Unknown | jayashreedevi2020 | 53 | facebook/wav2vec2-xls-r-300m | 64.7 |
| whisper-small-khmer | medium | Wer | Unknown | seanghay | 54 | — | 64.64 |
| moonshine-tiny-uk-ONNX | low | Fleurs (WER) ↓ | moonshine-tiny-uk | onnx-community | 55 | UsefulSensors/moonshine-tiny-uk | 63.83 |
| moonshine-tiny-uk | low | Fleurs (WER) ↓ | moonshine-tiny-uk | UsefulSensors | 56 | UsefulSensors/moonshine-tiny-uk | 63.83 |
| whisper-small-khmer-v2 | medium | Wer | Unknown | seanghay | 57 | — | 61.65 |
| swahili | medium | Wer | Unknown | etonkou | 58 | facebook/wav2vec2-base | 60.55 |
| wav2vec2-large-xls-r-300m-marathi-cv8 | medium | Wer | Unknown | infinitejoy | 59 | — | 60.49 |
| whisper-large-v3-turbo-arabic-dialectal | low | Zero-shot WER | whisper-large-v3-turbo | oddadmix | 60 | openai/whisper-large-v3-turbo | 59 |
| cohere-transcribe-arabic-07-2026-dialectal | low | Zero-shot WER | cohere-transcribe-arabic-07-2026 | oddadmix | 61 | CohereLabs/cohere-transcribe-arabic-07-2026 | 59 |
| english-filipino-wav2vec2-l-xls-r-test-09 | medium | Wer | Unknown | Khalsuu | 62 | — | 57.5 |
| afrivoices-asr-baobab | low | WER | Unknown | Okpeyemi | 63 | — | 56.6 |
| monsoon-whisper-medium-gigaspeech2 | high | WER (GS2) | Unknown | typhoon-ai | 64 | — | 55.64 |
| qwen3-asr-0.6b-arabic-dialectal-v2 | low | WER | qwen3-asr-0-6b-hf | oddadmix | 65 | Qwen/Qwen3-ASR-0.6B-hf | 55.14 |
| whisper-large-v3-myanmar | medium | Wer | whisper-large-v3 | chuuhtetnaing | 66 | openai/whisper-large-v3 | 54.9 |
| whisper_small_ps_augmented | medium | Wer | Unknown | ihanif | 67 | ihanif/whisper_small_ps_augmented | 53.62 |
| asr-cherokee | medium | Wer | Unknown | charliemcvicker | 68 | facebook/wav2vec2-large-xlsr-53 | 53.01 |
| whisper-small-p0km | medium | Wer | whisper-small | rick-c831 | 69 | openai/whisper-small | 52.73 |
| azerbaijani-whisper-turbo | low | WER | whisper-large-v3-turbo | LocalDoc | 70 | openai/whisper-large-v3-turbo | 52.17 |
| azerbaijani-whisper-small | high | WER | whisper-small | LocalDoc | 71 | openai/whisper-small | 52.17 |
| wav2vec2-xls-r-300m-pashto | medium | Wer | Unknown | ihanif | 72 | facebook/wav2vec2-xls-r-300m | 51.59 |
| whisper-tiny-hausa1 | medium | Wer | Unknown | EYEDOL | 73 | EYEDOL/whisper-tiny-hausa | 50.24 |
| cdli-whisper-ml-lug-swa-sunbird-full-a40 | medium | Wer | asr-whisper-large-v3-salt | KasuleTrevor | 74 | Sunbird/asr-whisper-large-v3-salt | 50.22 |
| whisper-small-shona | medium | Wer | whisper-small | steja | 75 | openai/whisper-small | 49.91 |
| whisper-large-v3-turbo-lt | low | WER | whisper-large-v3-turbo | svogunas | 76 | openai/whisper-large-v3-turbo | 49.66 |
| wav2vec2-xls-r-300m-sk-cv8 | medium | WER | Unknown | comodoro | 77 | — | 49.57 |
| mms-300m-waxal-mas | medium | **Test WER** | Unknown | waxal-benchmarking | 78 | facebook/mms-300m | 49.5 |
| mms-300m-waxal-mas | medium | **WER** | Unknown | waxal-benchmarking | 79 | facebook/mms-300m | 49.5 |
| wav2vec2-large-xlsr-53-punjabi | medium | Wer | Unknown | kingabzpro | 80 | Harveenchadha/vakyansh-wav2vec2-punjabi-pam-10 | 49.39 |
| whisper-small-myanmar | medium | Wer | whisper-small | chuuhtetnaing | 81 | openai/whisper-small | 49.06 |
| whisper-tiny-waxal-lin | medium | **Test WER** | whisper-tiny | waxal-benchmarking | 82 | openai/whisper-tiny | 49 |
| whisper-tiny-waxal-lin | medium | **WER** | whisper-tiny | waxal-benchmarking | 83 | openai/whisper-tiny | 49 |
| whisper-small-darija | medium | Wer | whisper-small | ychafiqui | 84 | openai/whisper-small | 48.27 |
| whisper-tiny-hausa2 | medium | Wer | Unknown | EYEDOL | 85 | EYEDOL/whisper-tiny-hausa1 | 48.09 |
| whisper-tiny-mla-24lang | high | WER% | whisper-tiny | burakaydinofficial | 86 | openai/whisper-tiny | 47.6 |
| Sunflower-Gemma4-E2B | low | Macro Language WER | gemma-4-e2b | Sunbird | 87 | google/gemma-4-E2B-it | 47.2 |
| whisper-small-hu | low | WER | whisper-small | SubZtep | 88 | openai/whisper-small | 46.79 |
| wav2vec2-large-xls-r-300m-bulgarian | medium | Wer | Unknown | infinitejoy | 89 | — | 46.74 |
| whisper-small-sinhala_v3 | medium | Wer | whisper-small | Lingalingeswaran | 90 | openai/whisper-small | 46.46 |
| whisper-tiny-waxal-lin | low | Wer | whisper-tiny | waxal-benchmarking | 91 | openai/whisper-tiny | 46.13 |
| whisper-base-bn | medium | WER | whisper-base | shhossain | 92 | openai/whisper-base | 46 |
| cohere-transcribe-arabic-07-2026-int8 | low | WER vs. full fp16 | cohere-transcribe-arabic-07-2026 | NAMAA-Space | 93 | CohereLabs/cohere-transcribe-arabic-07-2026 | 46 |
| cdli-whisper-ml-lug-swa-sunbird-full-a40-lr1e4 | medium | Wer | asr-whisper-large-v3-salt | KasuleTrevor | 94 | Sunbird/asr-whisper-large-v3-salt | 45.48 |
| whisper-tiny-oilgas-russian | medium | WER | whisper-tiny | N07P | 95 | openai/whisper-tiny | 45.08 |
| whisper-small-waxal-ful | low | Wer | whisper-small | waxal-benchmarking | 96 | openai/whisper-small | 44.82 |
| whisper-small.en-atco2-asr | medium | Wer | Unknown | jlvdoorn | 97 | openai/whisper-small.en | 44.66 |
| whisper-tiny.en | medium | Wer | Unknown | eai6 | 98 | openai/whisper-tiny.en | 44.35 |
| whisper-wolof | medium | Wer | whisper-small | cibfaye | 99 | openai/whisper-small | 43.94 |
| STT-Whisper-Pashto | medium | WER | whisper-small | Sabtain-Dev | 100 | openai/whisper-small | 43.83 |
| cdli-whisper-ml-lug-swa-sunbird-full-a40-lr5e5-5k | medium | Wer | asr-whisper-large-v3-salt | KasuleTrevor | 101 | Sunbird/asr-whisper-large-v3-salt | 43.44 |
| wav2vec2-large-mms-1b-somalia | medium | Wer | mms-1b-all | skydheere | 102 | facebook/mms-1b-all | 43.3 |
| whisper-small-arabic-dialectal | high | Zero-shot WER | whisper-small | oddadmix | 103 | openai/whisper-small | 42.8 |
| whisper-small-waxal-lin | medium | **Test WER** | whisper-small | waxal-benchmarking | 104 | openai/whisper-small | 42.7 |
| whisper-small-waxal-lin | medium | **WER** | whisper-small | waxal-benchmarking | 105 | openai/whisper-small | 42.7 |
| mms-300m-waxal-lin | medium | **Test WER** | Unknown | waxal-benchmarking | 106 | facebook/mms-300m | 42.6 |
| mms-300m-waxal-lin | medium | **WER** | Unknown | waxal-benchmarking | 107 | facebook/mms-300m | 42.6 |
| whisper-small-waxal-ful | medium | **Test WER** | whisper-small | waxal-benchmarking | 108 | openai/whisper-small | 42.6 |
| whisper-small-waxal-ful | medium | **WER** | whisper-small | waxal-benchmarking | 109 | openai/whisper-small | 42.6 |
| Ethio-ASR-tigrinya | low | Tigrinya WER (↓) | w2v-bert-2-0 | badrex | 110 | facebook/w2v-bert-2.0 | 42.42 |
| mms-300m-waxal-ach | medium | **Test WER** | Unknown | waxal-benchmarking | 111 | facebook/mms-300m | 42.3 |
| mms-300m-waxal-ach | medium | **WER** | Unknown | waxal-benchmarking | 112 | facebook/mms-300m | 42.3 |
| whisper-small-zh_tw | medium | Wer | Unknown | Jingmiao | 113 | — | 41.97 |
| griot-nano-1 | low | WER | Unknown | Qlerqly | 114 | — | 41.67 |
| griot-nano-1-onnx | low | WER | Unknown | Qlerqly | 115 | — | 41.67 |
| wav2vec2-large-xlsr-53-finnish | high | WER | Unknown | jonatasgrosman | 116 | — | 41.6 |
| BlueMagpie-TTS | low | WER | Unknown | OpenFormosa | 117 | OpenFormosa/BlueMagpie-TTS | 41.48 |
| whisper-small-he | medium | Wer | whisper-small | tomtom5 | 118 | openai/whisper-small | 41.34 |
| whisper-large-v3-turbo-swiss-german | low | WER | whisper-large-v3-turbo | Flurin17 | 119 | openai/whisper-large-v3-turbo | 41.05 |
| whisper-small-arabic-dialectal-v2 | low | WER | whisper-small | oddadmix | 120 | openai/whisper-small | 40.34 |
| VoiceTut-TTS | medium | WER — Egyptian Arabic | omnivoice | mohammedaly22 | 121 | k2-fsa/OmniVoice | 40 |
| whisper-small-waxal-lin | low | Wer | whisper-small | waxal-benchmarking | 122 | openai/whisper-small | 39.94 |
| wav2vec2-base-it-latin | medium | WER | Unknown | lsb | 123 | facebook/wav2vec2-base | 39.8 |
| cdli-whisper-ml-ugeng-lug-swa-sunbird-full-a40-lr1e4-5k | medium | Wer | asr-whisper-large-v3-salt | KasuleTrevor | 124 | Sunbird/asr-whisper-large-v3-salt | 39.61 |
| wav2vec2-large-xlsr-53-arabic | high | WER | Unknown | jonatasgrosman | 125 | jonatasgrosman/wav2vec2-large-xlsr-53-arabic | 39.59 |
| whisper-medium-ml | medium | WER | whisper-medium | thennal | 126 | openai/whisper-medium | 38.62 |
| mms-300m-waxal-nyn | medium | **Test WER** | Unknown | waxal-benchmarking | 127 | facebook/mms-300m | 38.6 |
| mms-300m-waxal-nyn | medium | **WER** | Unknown | waxal-benchmarking | 128 | facebook/mms-300m | 38.6 |
| xls-r-uzbek-cv8 | medium | Wer | Unknown | lucio | 129 | facebook/wav2vec2-xls-r-300m | 38.52 |
| cdli-whisper-ml-ugeng-lug-swa-sunbird-full-a40-lr5e5 | medium | Wer | asr-whisper-large-v3-salt | KasuleTrevor | 130 | Sunbird/asr-whisper-large-v3-salt | 37.97 |
| wav2vec2-child-en-tokenizer-4 | medium | Wer | Unknown | jaeyeon | 131 | — | 37.69 |
| whisper-small-malayalam | medium | WER | whisper-small | sajilck | 132 | openai/whisper-small | 37.64 |
| my_zh_CN_asr_cv13_model | medium | Wer | Unknown | tristayqc | 133 | jonatasgrosman/wav2vec2-large-xlsr-53-chinese-zh-cn | 37.5 |
| wav2vec2-base-timit-ms | medium | Wer | Unknown | saribalgar | 134 | facebook/wav2vec2-base | 37.22 |
| whisper-base-mla-24lang | high | WER% | whisper-base | burakaydinofficial | 135 | openai/whisper-base | 37 |
| whisper-small-waxal-sna | low | Wer | whisper-small | waxal-benchmarking | 136 | openai/whisper-small | 36.3 |
| mms-1b-all-indic_voices-normalized-30_70-lg-hin-steps-10000-lr-0.0001-FT | medium | Wer | mms-1b-all | dianavdavidson | 137 | facebook/mms-1b-all | 36.09 |
| w2vbert-waxal-multi7 | medium | Wer | Unknown | sulaimank | 138 | sulaimank/w2v-bert-waxal | 35.88 |
| Wolof-HuBERT-CTC | medium | Wer | Unknown | soynade-research | 139 | soynade-research/Wolof-HuBERT-Base | 35.65 |
| w2v-bert-2.0-afrivoices-ea-6l-asr-int8-edge | medium | WER | w2v-bert-2-0 | leophill | 140 | leophill/w2v-bert-2.0-afrivoices-ea-6l-asr | 34.81 |
| wav2vec2-large-xlsr-lithuanian | medium | WER | Unknown | m3hrdadfi | 141 | — | 34.66 |
| whisper-small-uz-v1 | high | WER ↓ | whisper-small | OvozifyLabs | 142 | openai/whisper-small | 34.5 |
| whisper-small-combined-lin-sna-lug-olusola | medium | Wer | whisper-small | samolusola-personal | 143 | openai/whisper-small | 34.43 |
| whisper-large-v3-turbo-arabic-dialectal | medium | WER | whisper-large-v3-turbo | oddadmix | 144 | openai/whisper-large-v3-turbo | 34.4 |
| cohere-transcribe-arabic-07-2026-dialectal | medium | WER | cohere-transcribe-arabic-07-2026 | oddadmix | 145 | CohereLabs/cohere-transcribe-arabic-07-2026 | 34.4 |
| whisper-small-arabic-dialectal | medium | WER | whisper-small | oddadmix | 146 | openai/whisper-small | 34.4 |
| qwen3-asr-0.6b-arabic-dialectal | medium | WER | qwen3-asr-0-6b-hf | oddadmix | 147 | Qwen/Qwen3-ASR-0.6B-hf | 34.4 |
| whisper-medium-combined-lin-sna-lug-olusola | medium | Wer | whisper-medium | samolusola-personal | 148 | openai/whisper-medium | 34.36 |
| whisper-tiny-waxal-lug | medium | **Test WER** | whisper-tiny | waxal-benchmarking | 149 | openai/whisper-tiny | 33.8 |
| whisper-tiny-waxal-lug | medium | **WER** | whisper-tiny | waxal-benchmarking | 150 | openai/whisper-tiny | 33.8 |
| wav2vec2-xls-r-300m-hy | medium | Wer | Unknown | arampacha | 151 | — | 33.33 |
| whisper-large-v3-turbo-arabic-dialectal-v2 | low | WER | whisper-large-v3-turbo | oddadmix | 152 | openai/whisper-large-v3-turbo | 33.24 |
| whisper-base-armenian | medium | **Word Error Rate (WER)** | whisper-base | Chillarmo | 153 | openai/whisper-base | 33.19 |
| whisper-base-ru | medium | Wer | Unknown | ovsyannikovls | 154 | ovsyannikovls/whisper-base-ru | 33.1 |
| swahili-wav2vec2-asr | medium | WER | Unknown | RareElf | 155 | eddiegulay/wav2vec2-large-xlsr-mvc-swahili | 33 |
| mms-1b-zdj | medium | Wer | mms-1b-all | rifailabs | 156 | facebook/mms-1b-all | 32.96 |
| whisper-tiny-waxal-lug | low | Wer | whisper-tiny | waxal-benchmarking | 157 | openai/whisper-tiny | 32.87 |
| whisper-small-pashto | medium | Wer | Unknown | Adnan666 | 158 | ihanif/whisper_small_ps_augmented | 32.6 |
| whisper-tiny-en-minds14 | medium | Wer | whisper-tiny | nagorebravo | 159 | openai/whisper-tiny | 32.41 |
| pashto-whisper-small | medium | Wer | whisper-small | karimnaimy | 160 | openai/whisper-small | 32.34 |
| wav2vec2-large-xlsr-persian | medium | WER | Unknown | m3hrdadfi | 161 | — | 32.2 |
| Audar-ASR-V1-Flash | high | Open Universal Arabic ASR Leaderboard — Avg WER | Unknown | audarai | 162 | — | 32.04 |
| whisper-small-hi | medium | Wer | whisper-small | sanchit-gandhi | 163 | openai/whisper-small | 32.01 |
| whisper-large-v3-arabic-dialectal-v2 | low | WER | whisper-large-v3 | oddadmix | 164 | openai/whisper-large-v3 | 31.99 |
| whisper-small-persian-v1 | medium | Wer | whisper-small | aliyzd95 | 165 | openai/whisper-small | 31.93 |
| wav2vec2-large-xlsr-53-hungarian | high | WER | Unknown | jonatasgrosman | 166 | — | 31.4 |
| whisper-large-libyan | medium | Wer | whisper-large-v3 | Rhaodgh | 167 | openai/whisper-large-v3 | 31.36 |
| whisper-large-v3-turbo-arabic | medium | Wer | whisper-large-v3-turbo | mboushaba | 168 | openai/whisper-large-v3-turbo | 31.15 |
| Ethio-ASR-amharic | low | Amharic WER (↓) | w2v-bert-2-0 | badrex | 169 | facebook/w2v-bert-2.0 | 30.95 |
| whisper-large-v3-bn-cv17 | medium | Wer | whisper-large-v3 | bijaykumarsingh | 170 | openai/whisper-large-v3 | 30.84 |
| wav2vec2-large-xlsr-japanese | medium | WER | Unknown | vumichien | 171 | — | 30.84 |
| ASR-Somali | medium | Wer | Unknown | Mustafaa4a | 172 | — | 30.6 |
| sushrota-sanskrit-asr | high | In-the-wild user audio (327 clips) — WER | Unknown | prathoshap | 173 | — | 30.4 |
| code-switched-egyptian-arabic-whisper-small | low | **Baseline Whisper Small WER (%)** | whisper-small | IbrahimAmin | 174 | openai/whisper-small | 30.29 |
| wav2vec2-large-xlsr-53-persian | high | WER | Unknown | jonatasgrosman | 175 | — | 30.12 |
| wav2vec2-large-xls-r-300m-bg | medium | Wer | Unknown | anuragshas | 176 | — | 30.02 |
| uk_wav2vec2_with_stress_mark | medium | Wer | Unknown | mouseyy | 177 | mouseyy/result_data_2-3 | 29.84 |
| nemotron-3.5-swahili-streaming-asr | low | WER | nemotron-3-5-asr-streaming-0-6b | Tonykip | 178 | nvidia/nemotron-3.5-asr-streaming-0.6b | 29.8 |
| wav2vec2-xls-r-300m-korean | low | WER | Unknown | w11wo | 179 | — | 29.54 |
| filipino-wav2vec2-l-xls-r-300m-official | medium | Wer | Unknown | Khalsuu | 180 | — | 29.22 |
| whisper-small-si | medium | Wer | whisper-small | seniruk | 181 | openai/whisper-small | 29.14 |
| wav2vec2-large-xls-r-300m-bg-d2 | medium | Wer | Unknown | DrishtiSharma | 182 | — | 28.6 |
| gha-whisper-small-twi-v3 | medium | Wer | whisper-small | teckedd | 183 | teckedd/whisper-small-waxal-round2-specaug-v1 | 28.07 |
| xlsr-waxal-from-luganda | medium | Wer | Unknown | sulaimank | 184 | sulaimank/xls-r-300m-luganda-313hr | 28.06 |
| whisper-small-yoruba | low | WER (%) | whisper-small | LyngualLabs | 185 | openai/whisper-small | 27.92 |
| pashto-voice-v2 | medium | Wer | whisper-small | karimnaimy | 186 | openai/whisper-small | 27.7 |
| wav2vec2-large-xlsr-turkish | medium | WER | Unknown | m3hrdadfi | 187 | — | 27.51 |
| whisper-large-v2-malayalam | medium | Wer | Unknown | DrishtiSharma | 188 | — | 27.46 |
| egyptian-arabic-wav2vec2-xlsr-53 | low | **WER (%)** | Unknown | IbrahimAmin | 189 | omarxadel/wav2vec2-large-xlsr-53-arabic-egyptian | 27.2 |
| moonshine-tiny-ar-gguf | low | WER (FLEURS ar test) | moonshine-tiny-ar | handy-computer | 190 | UsefulSensors/moonshine-tiny-ar | 27.11 |
| wav2vec2-large-xls-r-300m-welsh | medium | Wer | Unknown | infinitejoy | 191 | — | 27.02 |
| wav2vec2-large-mms-1b-all-lingala-ojpl | medium | Wer | mms-1b-all | BrainTheos | 192 | facebook/mms-1b-all | 26.98 |
| whisper-medium-urdu | medium | Wer | Unknown | ihanif | 193 | — | 26.98 |
| whisper-small-waxal-sna | medium | **Test WER** | whisper-small | waxal-benchmarking | 194 | openai/whisper-small | 26.9 |
| whisper-small-waxal-sna | medium | **WER** | whisper-small | waxal-benchmarking | 195 | openai/whisper-small | 26.9 |
| whisper-meduim-mongolian | medium | Wer | whisper-medium | Cafet | 196 | openai/whisper-medium | 26.87 |
| whisper-large-v3-persian | medium | Wer | whisper-large-v3 | nezamisafa | 197 | openai/whisper-large-v3 | 26.54 |
| whisper-large-persian | medium | Wer | Unknown | steja | 198 | — | 26.37 |
| wav2vec2-large-mms-1b-azerbaijani-common_voice15.0 | medium | Wer | mms-1b-all | nijatzeynalov | 199 | facebook/mms-1b-all | 26.32 |
| whisper-large-v3-turbo-korean | medium | WER | whisper-large-v3-turbo | ghost613 | 200 | openai/whisper-large-v3-turbo | 26.26 |
| whisper-large-v3-turbo-urdu | medium | **WER** | whisper-large-v3-turbo | kingabzpro | 201 | openai/whisper-large-v3-turbo | 26.23 |
| whisper-large-v3-urdu | medium | **WER** | whisper-large-v3 | kingabzpro | 202 | openai/whisper-large-v3 | 26.02 |
| whisper-large-v3-eu | low | Wer | whisper-large-v3 | xezpeleta | 203 | openai/whisper-large-v3 | 25.89 |
| whisper-small-persian | medium | Wer | whisper-small | AmirMohseni | 204 | openai/whisper-small | 25.82 |
| whisper-large-v3-turbo-urdu | medium | Wer | whisper-large-v3-turbo | kingabzpro | 205 | openai/whisper-large-v3-turbo | 25.78 |
| whisper-small-yoruba-07-17 | medium | Wer | whisper-small | ccibeekeoc42 | 206 | openai/whisper-small | 25.68 |
| flix-swissgerman-full | medium | WER (measured) | Unknown | Flix-AI | 207 | — | 25.6 |
| w2v-bert-2.0-khmer | medium | Wer | w2v-bert-2-0 | Prakmlis | 208 | facebook/w2v-bert-2.0 | 25.59 |
| whisper-base-bungoma.en | medium | Wer | Unknown | eai6 | 209 | openai/whisper-base.en | 25.28 |
| tamil-asr-qwen3 | low | WER ↓ | qwen3-asr-1-7b | osmapi | 210 | Qwen/Qwen3-ASR-1.7B | 25.27 |
| nemotron-3.5-asr-streaming-slovenian | low | WER | nemotron-3-5-asr-streaming-0-6b | texdata | 211 | nvidia/nemotron-3.5-asr-streaming-0.6b | 25.07 |
| mms-300m-waxal-sna | medium | **Test WER** | Unknown | waxal-benchmarking | 212 | facebook/mms-300m | 25 |
| mms-300m-waxal-sna | medium | **WER** | Unknown | waxal-benchmarking | 213 | facebook/mms-300m | 25 |
| wav2vec2-xlsr-khmer | medium | WER | Unknown | gagan3012 | 214 | — | 24.96 |
| whisper-large-v3-turbo-darija | medium | WER | whisper-large-v3-turbo | anaszil | 215 | openai/whisper-large-v3-turbo | 24.9 |
| whisper-large-v3-turbo-darija | medium | **Word Error Rate (WER)** | whisper-large-v3-turbo | anaszil | 216 | openai/whisper-large-v3-turbo | 24.88 |
| wav2vec2-large-xlsr-japanese-hiragana | medium | WER | Unknown | vumichien | 217 | — | 24.74 |
| moonshine_tiny_pt_v05 | medium | Wer | Unknown | aomocelin | 218 | aomocelin/moonshine_tiny_pt_v04 | 24.74 |
| whisper-large-v3-vaani-hindi | high | WER | whisper-small | ARTPARK-IISc | 219 | openai/whisper-small | 24.66 |
| whisper-large-v3-turbo-russian-codeswitch | medium | WER | whisper-large-v3-turbo | coriollon | 220 | coriollon/whisper-large-v3-turbo-russian | 24.64 |
| whisper-large-v2-vietnamese-lyrics-transcription | high | **WER (%)<br>Case-Sensitive** | whisper-large-v2 | kelvinbksoh | 221 | openai/whisper-large-v2 | 24.61 |
| wav2vec2-bloom-speech-tgl | medium | Wer | Unknown | sil-ai | 222 | — | 24.57 |
| moonshine-base-ar-gguf | low | WER (FLEURS ar test) | Unknown | handy-computer | 223 | UsefulSensors/moonshine-base-ar | 24.45 |
| Cohere-Speech-Tashkeel-2B | medium | WER — fully diacritized | cohere-transcribe-arabic-07-2026 | NAMAA-Space | 224 | CohereLabs/cohere-transcribe-arabic-07-2026 | 24.35 |
| whisper-small-mla-24lang | high | WER% | whisper-small | burakaydinofficial | 225 | openai/whisper-small | 24.2 |
| openai_whisper-large-v3-turbo-NVFP4 | low | WER ↓ | whisper-large-v3-turbo | Reza2kn | 226 | openai/whisper-large-v3-turbo | 24.18 |
| Medical-Whisper-Large-v3 | low | WER | whisper-large-v3 | Na0s | 227 | openai/whisper-large-v3 | 24 |
| whisper-large-v3_finetuned_ugandan_english_nonstandard_speech_v1.0 | low | Standard WER | Unknown | cdli | 228 | — | 24 |
| whisper-large-v3-swahili-afrivoice-cv | medium | WER | whisper-large-v3 | ElizabethMwangi | 229 | openai/whisper-large-v3 | 23.93 |
| whisper-large-v3-afrikaans | low | WER (%) | whisper-large-v3 | andreoosthuizen | 230 | openai/whisper-large-v3 | 23.92 |
| whisper-large-v3-turbo-nepali | medium | Wer | whisper-large-v3-turbo | kiranpantha | 231 | openai/whisper-large-v3-turbo | 23.63 |
| whosper-large-v2 | medium | WER | whisper-large-v2 | CAYTU | 232 | openai/whisper-large-v2 | 23.45 |
| Audar-ASR-V1-Turbo | high | Open Universal Arabic ASR Leaderboard — Avg WER | Unknown | audarai | 233 | — | 23.17 |
| whisper_small-fa_v03 | medium | Wer | whisper-small | mohammadh128 | 234 | openai/whisper-small | 23.15 |
| parakeet-tdt-0.6b-EN-Medical | medium | WER | parakeet-tdt-0-6b-v3 | yuriyvnv | 235 | nvidia/parakeet-tdt-0.6b-v3 | 22.84 |
| whisper-small-egyptian-arabic | medium | WER | Unknown | MAdel121 | 236 | — | 22.79 |
| whisper-small-hy-2 | medium | Wer | whisper-small | ErikMkrtchyan | 237 | openai/whisper-small | 22.79 |
| whisper-small-egyptian-arabic | medium | WER | Unknown | MAdel121 | 238 | — | 22.69 |
| wav2vec2-base-760-turkish | medium | WER | Unknown | ceyda | 239 | facebook/wav2vec2-base | 22.6 |
| asr-wav2vec2-commonvoice-14-pt | low | Test WER | Unknown | speechbrain | 240 | — | 22.41 |
| whisper-medium-full-clinical-r16 | medium | Wer | whisper-medium | robello2 | 241 | openai/whisper-medium | 22.4 |
| whisper-base-google-fleurs-pt-br | medium | Wer | whisper-base | thiagobarbosa | 242 | openai/whisper-base | 22.3 |
| whisper-large-v3-atco2-asr-atcosim | medium | Wer | whisper-large-v3 | jlvdoorn | 243 | openai/whisper-large-v3 | 22.27 |
| whisper-large-v3-turbo-hindi-lora | medium | WER | whisper-large-v3-turbo | Tachyeon | 244 | openai/whisper-large-v3-turbo | 22.25 |
| wav2vec2-large-xls-r-300m-Urdu | high | WER | Unknown | kingabzpro | 245 | facebook/wav2vec2-xls-r-300m | 22.22 |
| whisper-small-waxal-lug | low | Wer | whisper-small | waxal-benchmarking | 246 | openai/whisper-small | 22.15 |
| uz-whisper-small-stt-v2 | high | Mozilla Common Voice 26 Uzbek test split — Test WER | whisper-small | maqsudxo1ja | 247 | maqsudxo1ja/uz-whisper-small-stt | 22.138 |
| whisper-small-mn | medium | WER | whisper-small | Blgn94 | 248 | openai/whisper-small | 22.1 |
| whisper-large-fa-v1 | low | WER (%) | whisper-large-v3-turbo | vhdm | 249 | openai/whisper-large-v3-turbo | 22.07 |
| moonshine-tiny-fr-ONNX | low | Word Error Rate (WER) | Unknown | onnx-community | 250 | Cornebidouil/moonshine-tiny-fr | 21.8 |
| moonshine-tiny-fr-ONNX | medium | **Word Error Rate (WER)** | Unknown | onnx-community | 251 | Cornebidouil/moonshine-tiny-fr | 21.8 |
| wav2vec2-xls-r-300m-bengali | medium | WER | Unknown | arijitx | 252 | — | 21.72 |
| whisper-small-waxal-lug | medium | **Test WER** | whisper-small | waxal-benchmarking | 253 | openai/whisper-small | 21.6 |
| whisper-small-waxal-lug | medium | **WER** | whisper-small | waxal-benchmarking | 254 | openai/whisper-small | 21.6 |
| whisper-large-v3-urdu | medium | Wer | whisper-large-v3 | kingabzpro | 255 | openai/whisper-large-v3 | 21.47 |
| xlsr-luganda-waxal | medium | Wer | Unknown | sulaimank | 256 | sulaimank/wav2vec2-xlsr-swahili-400hr | 21.44 |
| wav2vec2-xls-r-300m-wolof-lm | medium | Wer | Unknown | abdouaziiz | 257 | — | 21.25 |
| whisper-tiny_ro-80mel | medium | WER | whisper-tiny | IonGrozea | 258 | openai/whisper-tiny | 21.2 |
| w2v-bert-2.0-khmer-v4 | medium | Wer | w2v-bert-2-0 | Prakmlis | 259 | facebook/w2v-bert-2.0 | 21.16 |
| whisper-malagasy-medium | medium | WER | whisper-medium | Flo976 | 260 | openai/whisper-medium | 20.78 |
| whisper-small-yoruba | medium | **Test WER** | whisper-small | LyngualLabs | 261 | openai/whisper-small | 20.76 |
| whisper-tiny_ro-80mel | medium | WER (normalized: locase, punctuation stripped) | whisper-tiny | IonGrozea | 262 | openai/whisper-tiny | 20.73 |
| wav2vec2-xls-r-300m-ftspeech | low | **WER without LM** | Unknown | saattrupdan | 263 | facebook/wav2vec2-xls-r-300m | 20.48 |
| whisper-large-v3-high-mixed-nl | medium | **Test WER (MLS)** | whisper-large-v3 | yuriyvnv | 264 | openai/whisper-large-v3 | 20.29 |
| w2v-xls-r-uk | medium | WER | Unknown | Yehor | 265 | facebook/wav2vec2-xls-r-300m | 20.24 |
| audioX-south-v1 | low | Avg WER | Unknown | jiviai | 266 | — | 20.1 |
| w2vbert-luganda-waxal-punct-v2 | low | Wer Keep | Unknown | sulaimank | 267 | sulaimank/w2vbert-luganda-waxal-stage2 | 20.07 |
| whisper-large-v2-mn-13 | medium | Wer | Unknown | bayartsogt | 268 | — | 20.02 |
| Omnilingual-ASR-CTC-300M-CoreML-INT8-10s | low | WER | Unknown | aufklarer | 269 | facebook/omniASR-CTC-300M | 20 |
| Omnilingual-ASR-CTC-300M-MLX-4bit | low | WER | Unknown | aufklarer | 270 | facebook/omniASR-CTC-300M | 20 |
| whisper-medium-zulu | medium | Wer | whisper-medium | TheirStory | 271 | openai/whisper-medium | 19.93 |
| wav2vec2-large-mms-1b-hungarian | medium | WER | mms-1b-all | GaborMadarasz | 272 | facebook/mms-1b-all | 19.93 |
| Whisper-Large-v3-turbo-STT-Zeroth-KO-v2 | medium | Wer | whisper-large-v3-turbo | o0dimplz0o | 273 | openai/whisper-large-v3-turbo | 19.91 |
| QwenCleo-ASR | low | WER all | qwen3-asr-1-7b | mohammedaly22 | 274 | Qwen/Qwen3-ASR-1.7B | 19.85 |
| openai_whisper-large-v3-NVFP4 | low | WER ↓ | whisper-large-v3 | Reza2kn | 275 | openai/whisper-large-v3 | 19.74 |
| SraVaani-1.0 | low | Avg WER | Unknown | ARTPARK-IISc | 276 | — | 19.4 |
| Test | low | WER % | Unknown | AbelWa | 277 | — | 19.4 |
| parakeet-tdt-0.6b-EN-Medical | low | Zero-shot WER | Zero-shot CER | parakeet-tdt-0-6b-v3 | yuriyvnv | 278 | nvidia/parakeet-tdt-0.6b-v3 | 19.22 |
| cohere-transcribe-arabic-07-2026-dialectal-v2 | low | Fine-tuned WER | cohere-transcribe-arabic-07-2026 | oddadmix | 279 | CohereLabs/cohere-transcribe-arabic-07-2026 | 19.1 |
| whisper-small_finetuned_ghanian_ga_standard_speech_v1.0 | high | cdli/ghanian_ga_standard_speech_v1.0 test split — Standard corpus-level WER | Unknown | cdli | 280 | — | 19 |
| whisper-large-v3-turbo-turkish | medium | Wer | whisper-large-v3-turbo | selimc | 281 | openai/whisper-large-v3-turbo | 18.92 |
| moonshine-tiny-uk-gguf | low | WER (FLEURS uk test) | moonshine-tiny-uk | handy-computer | 282 | UsefulSensors/moonshine-tiny-uk | 18.83 |
| Kabyle_xlsr | medium | Wer | Unknown | Akashpb13 | 283 | — | 18.79 |
| whisper-large-v3-nepali | medium | Wer | whisper-large-v3 | kiranpantha | 284 | openai/whisper-large-v3 | 18.73 |
| w2v-bert-waxal | medium | Wer | Unknown | sulaimank | 285 | sulaimank/w2v-bert-lingala-109h | 18.67 |
| wav2vec2-large-xls-r-1b-Swedish | medium | Wer | Unknown | kingabzpro | 286 | facebook/wav2vec2-xls-r-1b | 18.44 |
| whisper-small-bn | medium | Wer | Unknown | anuragshas | 287 | — | 18.22 |
| parakeet-ctc-110m-coreml | medium | WER | Unknown | FluidInference | 288 | nvidia/parakeet-tdt_ctc-110m | 17.97 |
| whisper-small-uyghur-thugy20 | medium | Wer | whisper-small | ixxan | 289 | openai/whisper-small | 17.94 |
| medasr-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 290 | google/medasr | 17.88 |
| wav2vec2-xls-r-300m-hebrew | medium | WER | Unknown | imvladikon | 291 | facebook/wav2vec2-xls-r-300m | 17.73 |
| stt_uz_fastconformer_hybrid_large_pc | low | **WER(%)** | Unknown | nvidia | 292 | nvidia/stt_uz_fastconformer_hybrid_large_pc | 17.52 |
| xlsr-luganda-waxal-pseudo | medium | Wer | Unknown | sulaimank | 293 | — | 17.49 |
| w2v-bert-uk-v2.1 | medium | WER | w2v-bert-2-0 | Yehor | 294 | facebook/w2v-bert-2.0 | 17.34 |
| VibeVoice-ASR-HF | low | WER (%) | vibevoice-asr-hf | microsoft | 295 | microsoft/VibeVoice-ASR-HF | 17.2 |
| parakeet-tdt-1.1b-gguf | medium | % WER | Unknown | handy-computer | 296 | nvidia/parakeet-tdt-1.1b | 17.18 |
| parakeet-tdt-1.1b | medium | % WER | Unknown | nvidia | 297 | nvidia/parakeet-tdt-1.1b | 17.18 |
| whisper-large-v3-atco2-asr | medium | Wer | whisper-large-v3 | jlvdoorn | 298 | openai/whisper-large-v3 | 17.04 |
| canary-180m-flash-gguf | medium | % WER | Unknown | handy-computer | 299 | nvidia/canary-180m-flash | 16.92 |
| canary-180m-flash | medium | % WER | Unknown | nvidia | 300 | nvidia/canary-180m-flash | 16.92 |
| mms-300m-waxal-lug | medium | **Test WER** | Unknown | waxal-benchmarking | 301 | facebook/mms-300m | 16.9 |
| mms-300m-waxal-lug | medium | **WER** | Unknown | waxal-benchmarking | 302 | facebook/mms-300m | 16.9 |
| ne-asr | low | Test WER | Unknown | MWirelabs | 303 | — | 16.89 |
| wav2vec2-xls-r-300m-en-atc-uwb-atcc-and-atcosim | medium | Wer | Unknown | Jzuluaga | 304 | — | 16.87 |
| asr-wav2vec2-commonvoice-14-en | low | Test WER | Unknown | speechbrain | 305 | — | 16.86 |
| canary-qwen-2.5b | medium | % WER | canary-1b-flash | nvidia | 306 | nvidia/canary-1b-flash | 16.71 |
| canary-qwen-2.5b-gguf | medium | % WER | canary-qwen-2-5b | handy-computer | 307 | nvidia/canary-qwen-2.5b | 16.71 |
| uzbek_stt_v1 | low | WER | whisper-medium | Kotib | 308 | openai/whisper-medium | 16.7 |
| uzbek_stt_v1 | low | WER | whisper-medium | Koyiljon | 309 | openai/whisper-medium | 16.7 |
| uzbek_stt_v1 | low | WER | whisper-medium | Uzbekswe | 310 | openai/whisper-medium | 16.7 |
| stt_ar_fastconformer_hybrid_large_pcd_v1.0 | low | **MASC Test WER** | Unknown | nvidia | 311 | nvidia/stt_ar_fastconformer_hybrid_large_pcd_v1.0 | 16.67 |
| whisper-tiny_to_Chinese_accent_SP_PHaPS | medium | Wer | whisper-tiny | liuh6 | 312 | openai/whisper-tiny | 16.58 |
| whisper-large-v3-turbo-tj | medium | Wer | whisper-large-v3-turbo | re-skill | 313 | openai/whisper-large-v3-turbo | 16.43 |
| kiswahili_sahihi_asr_adapted_2 | low | WER (%) | whisper-medium | keystats | 314 | keystats/kiswahili_sahihi_asr | 16.23 |
| whisper-base-ar-quran-mix-norm | medium | Wer | whisper-base | deepdml | 315 | openai/whisper-base | 16.2 |
| wav2vec2-large-mms-1b-tatar | medium | Wer | mms-1b-all | AigizK | 316 | facebook/mms-1b-all | 16.18 |
| whisper-small-sr | high | CV WER | whisper-small | istomin9192 | 317 | openai/whisper-small | 16.14 |
| whisper-tiny-es-mix-norm | high | Final evaluation set — WER | whisper-tiny | deepdml | 318 | openai/whisper-tiny | 16.069 |
| surt-small-v3 | medium | WER | whisper-small | surindersinghssj | 319 | openai/whisper-small | 15.84 |
| whisper-base_ro-80mel | medium | WER | whisper-base | IonGrozea | 320 | openai/whisper-base | 15.58 |
| whisper-th-medium-timestamp | medium | WER | whisper-medium | biodatlab | 321 | biodatlab/whisper-th-medium-combined | 15.57 |
| whisper-th-large-combined | medium | Wer | Unknown | biodatlab | 322 | — | 15.51 |
| DiarizationLM-8b-Fisher-v2 | low | WER (%) | Unknown | 323 | — | 15.48 | |
| DiarizationLM-13b-Fisher-v1 | low | WER (%) | Unknown | 324 | — | 15.48 | |
| DiarizationLM-8b-Fisher-v1 | low | WER (%) | Unknown | 325 | — | 15.48 | |
| qwen3-asr-0.6b-hinglish | low | Srota WER | qwen3-asr-0-6b | moorlee | 326 | Qwen/Qwen3-ASR-0.6B | 15.41 |
| whisper-large-v3-turbo-armenian | medium | **Word Error Rate (WER)** | whisper-large-v3-turbo | Chillarmo | 327 | openai/whisper-large-v3-turbo | 15.31 |
| parakeet-tdt-0.6b-v3-int8 | medium | WER (TED-LIUM) | parakeet-tdt-0-6b-v3 | sonic-speech | 328 | sonic-speech/parakeet-tdt-0.6b-v3 | 15.1 |
| whisper-base_ro-80mel | medium | WER (normalized: locase, punctuation stripped) | whisper-base | IonGrozea | 329 | openai/whisper-base | 15.09 |
| za-anv-multilingual-whisper-v3-turbo | medium | Wer | whisper-large-v3-turbo | dsfsi-anv | 330 | openai/whisper-large-v3-turbo | 15.01 |
| whisper-large-v3-lingala-asr | medium | Wer | whisper-large-v3 | noirlab | 331 | openai/whisper-large-v3 | 14.83 |
| whisper-large-finnish-v3 | medium | WER | Unknown | Finnish-NLP | 332 | — | 14.81 |
| wav2vec2-xls-r-300m-cs-250 | medium | Wer | Unknown | comodoro | 333 | facebook/wav2vec2-xls-r-300m | 14.75 |
| NaijaVox-2.0 | low | WER | whisper-large-v3 | Axiveri | 334 | openai/whisper-large-v3 | 14.7 |
| whisper-small-pt-cv11-v5 | medium | Wer | whisper-small | jonatasgrosman | 335 | openai/whisper-small | 14.68 |
| canary-1b-flash-gguf | medium | % WER | canary-1b-flash | handy-computer | 336 | nvidia/canary-1b-flash | 14.66 |
| canary-1b-flash | medium | % WER | canary-1b-flash | nvidia | 337 | nvidia/canary-1b-flash | 14.66 |
| canary-1b-flash-gguf | medium | % WER | canary-1b-flash | memoravox | 338 | nvidia/canary-1b-flash | 14.66 |
| canary-1b-gguf | medium | % WER | Unknown | handy-computer | 339 | nvidia/canary-1b | 14.64 |
| canary-1b | medium | % WER | Unknown | nvidia | 340 | nvidia/canary-1b | 14.64 |
| wav2vec2-bert-CV16-en | medium | Wer | Unknown | hf-audio | 341 | ylacombe/w2v-bert-2.0 | 14.55 |
| moonshine-base-uk-gguf | low | WER (FLEURS uk test) | Unknown | handy-computer | 342 | UsefulSensors/moonshine-base-uk | 14.49 |
| qwen3-asr-arabic-ksa | medium | **WER** | qwen3-asr-1-7b | vadimbelsky | 343 | Qwen/Qwen3-ASR-1.7B | 14.41 |
| wav2vec2-base-japanese-asr | medium | WER | Unknown | TKU410410103 | 344 | facebook/wav2vec2-base | 14.18 |
| whisper-large-fa-v1 | medium | **Word Error Rate (WER)** | whisper-large-v3-turbo | vhdm | 345 | openai/whisper-large-v3-turbo | 14.07 |
| whisper-uz | medium | Wer | Unknown | jmshd | 346 | jamshidahmadov/whisper-uz | 14.01 |
| odia_whisper_small_v3.0 | medium | Wer | Unknown | Ranjit | 347 | — | 13.86 |
| whisper-medium-et | low | WER | Unknown | TalTechNLP | 348 | — | 13.8 |
| canary-1b-v2 | low | **WER ↓** | Unknown | nvidia | 349 | nvidia/canary-1b-v2 | 13.78 |
| tadabur-Whisper-Small | low | WER ↓ | whisper-small | FaisaI | 350 | openai/whisper-small | 13.78 |
| sauti-asr | medium | WER | whisper-large-v3-turbo | msingiai | 351 | microsoft/paza-whisper-large-v3-turbo | 13.72 |
| Nemotron-3.5-ASR-Streaming-0.6B-MLX-5bit | high | WER % | nemotron-3-5-asr-streaming-0-6b | aufklarer | 352 | nvidia/nemotron-3.5-asr-streaming-0.6b | 13.66 |
| whisper-large-v2-marathi | medium | Wer | Unknown | DrishtiSharma | 353 | — | 13.64 |
| IndicConformer-GGUF | medium | **IndicConformer GGUF (WER)** | Unknown | Singla0009 | 354 | ai4bharat/indicconformer_stt_hi_hybrid_rnnt_large | 13.5 |
| whisper-small-ja | medium | Wer | whisper-small | vumichien | 355 | openai/whisper-small | 13.47 |
| stt_kr_conformer_ctc_medium | medium | WER | Unknown | SungBeom | 356 | — | 13.45 |
| whisper-large-v2-bg | medium | Wer | Unknown | anuragshas | 357 | — | 13.4 |
| Nemotron-3.5-ASR-Streaming-0.6B-MLX-8bit | high | WER % | nemotron-3-5-asr-streaming-0-6b | aufklarer | 358 | nvidia/nemotron-3.5-asr-streaming-0.6b | 13.37 |
| whisper-large-v3-turbo-russian | medium | WER (aggregate) | whisper-large-v3-turbo | coriollon | 359 | openai/whisper-large-v3-turbo | 13.25 |
| wav2vec2-xls-r-latin | high | LLPSI 5-fold CV — Average Strict WER | Unknown | njand | 360 | facebook/wav2vec2-xls-r-300m | 13.2 |
| moonshine-tiny-vi-gguf | low | WER (FLEURS vi test) | moonshine-tiny-vi | handy-computer | 361 | UsefulSensors/moonshine-tiny-vi | 13.16 |
| whisper-th-small-combined | medium | WER | whisper-small | biodatlab | 362 | openai/whisper-small | 13.14 |
| whisper-medium-hindi-hinglish-asr-fine-tuned | high | WER | whisper-medium | shujaAK | 363 | openai/whisper-medium | 13.01 |
| parakeet-tdt_ctc-1.1b-gguf | medium | % WER | Unknown | handy-computer | 364 | nvidia/parakeet-tdt_ctc-1.1b | 12.81 |
| parakeet-tdt_ctc-1.1b | medium | % WER | Unknown | nvidia | 365 | nvidia/parakeet-tdt_ctc-1.1b | 12.81 |
| w2vbert-luganda-waxal | medium | Wer | w2v-bert-2-0 | sulaimank | 366 | sulaimank/w2v-bert-2.0-lg-CV-Fleurs-300 | 12.78 |
| wav2vec2-large-xlsr-german | medium | WER | Unknown | maxidl | 367 | — | 12.77 |
| w2vbert-shona-waxal-punct-v2 | low | Wer Keep | w2v-bert-2-0 | sulaimank | 368 | sulaimank/w2vbert-shona-waxal | 12.76 |
| parakeet-tdt-0.6b-v3-ptBR-gguf | low | WER | Unknown | luizhrocha | 369 | alexandreacff/parakeet-tdt-0.6b-v3-ptBR-plus | 12.7 |
| whisper-base-es-mix-norm | high | Final evaluation set — WER | whisper-base | deepdml | 370 | openai/whisper-base | 12.406 |
| nb-wav2vec2-300m-nynorsk | medium | WER | Unknown | NbAiLab | 371 | — | 12.22 |
| whisper-large-v3-turkish-general | low | WER | whisper-large-v3 | turkmedstt | 372 | openai/whisper-large-v3 | 12.13 |
| wav2vec2-base-german-cv9 | medium | Wer | Unknown | oliverguhr | 373 | facebook/wav2vec2-base | 12.09 |
| nemotron-ocr-v1 | medium | Bag-of-word Error Rate | Unknown | nvidia | 374 | nvidia/nemotron-ocr-v1 | 12.03 |
| parakeet-tdt-0.6b-v3-gguf | medium | **Average WER ↓** | parakeet-tdt-0-6b-v3 | handy-computer | 375 | nvidia/parakeet-tdt-0.6b-v3 | 11.97 |
| parakeet-tdt-0.6b-v3 | medium | **Average WER ↓** | parakeet-tdt-0-6b-v3 | nvidia | 376 | nvidia/parakeet-tdt-0.6b-v3 | 11.97 |
| kazakh-whisper-large-v3-turbo | low | WER ↓ | whisper-large-v3-turbo | shyngys879 | 377 | openai/whisper-large-v3-turbo | 11.8 |
| hviske-v5 | low | WER | Unknown | syvai | 378 | — | 11.8 |
| romanian-wav2vec2 | medium | Wer | Unknown | gigant | 379 | facebook/wav2vec2-xls-r-300m | 11.74 |
| Nemotron-Speech-Streaming-0.6B-CoreML-INT8 | high | WER (%) | nemotron-speech-streaming-en-0-6b | aufklarer | 380 | nvidia/nemotron-speech-streaming-en-0.6b | 11.73 |
| wav2vec2-large-xlsr-53-greek | high | WER | Unknown | jonatasgrosman | 381 | — | 11.62 |
| stt_ar_fastconformer_hybrid_large_pc_v1.0 | low | **MASC Test WER** | Unknown | nvidia | 382 | nvidia/stt_ar_fastconformer_hybrid_large_pc_v1.0 | 11.46 |
| moonshine-tiny-de-fidoriel-GGUF | medium | WER | Unknown | cstr | 383 | fidoriel/moonshine-tiny-de | 11.4 |
| nb-wav2vec2-1b-nynorsk | medium | WER | Unknown | NbAiLab | 384 | — | 11.32 |
| Zipformer-cr-ctc-transducer-XL-290M | low | WER | Unknown | soundsgoodai | 385 | — | 11.3 |
| Shrutimala_Bangla_ASR | medium | WER | w2v-bert-2-0 | sazzadul | 386 | facebook/w2v-bert-2.0 | 11.26 |
| whisper-large-v2-bn | medium | Wer | Unknown | anuragshas | 387 | — | 11.11 |
| omi-med-stt-v1 | low | WER | Unknown | omi-health | 388 | nvidia/parakeet-tdt-0.6b-v2 | 11.1 |
| omi-med-stt-v1-mlx-q8 | low | WER | Unknown | omi-health | 389 | omi-health/omi-med-stt-v1 | 11.1 |
| kiswahili_sahihi_asr_adapted_2 | medium | **Final Validation WER** | whisper-medium | keystats | 390 | keystats/kiswahili_sahihi_asr | 11.09 |
| whisper-large-v3-turbo-greek-greece | medium | Wer | whisper-large-v3-turbo | sam8000 | 391 | openai/whisper-large-v3-turbo | 11.06 |
| cohere-transcribe-arabic-07-2026-gguf | low | WER (FLEURS Arabic test) | cohere-transcribe-arabic-07-2026 | handy-computer | 392 | CohereLabs/cohere-transcribe-arabic-07-2026 | 11.02 |
| Qwen3-ASR-1.7B-nvfp4 | high | WER | qwen3-asr-1-7b | vrfai | 393 | Qwen/Qwen3-ASR-1.7B | 10.73 |
| parakeet-rnnt-110m-da-dk | low | **WER (%)** | Unknown | nvidia | 394 | nvidia/parakeet-rnnt-110m-da-dk | 10.7 |
| wav2vec2-xls-r-300m-cv8-turkish | low | WER | Unknown | mpoyraz | 395 | — | 10.61 |
| wav2vec2-large-xlsr-53-greek | medium | WER on CommonVoice Test (%) &ast; | Unknown | lighteternal | 396 | — | 10.5 |
| wav2vec2-large-960h-lv60-self-en-atc-uwb-atcc-and-atcosim | medium | Wer | Unknown | Jzuluaga | 397 | facebook/wav2vec2-large-960h-lv60-self | 10.49 |
| wav2vec2-xls-r-parlaspeech-hr | low | WER | Unknown | classla | 398 | — | 10.46 |
| parakeet-tdt-0.6b-v3-fr-tv-media | low | Base WER | parakeet-tdt-0-6b-v3 | Archime | 399 | nvidia/parakeet-tdt-0.6b-v3 | 10.42 |
| wav2vec2-large-xlsr-persian-v3 | medium | WER | Unknown | m3hrdadfi | 400 | — | 10.36 |
| hviske-v5.3 | low | raw WER | Unknown | syvai | 401 | syvai/hviske-v5.1 | 10.26 |
| higgs-audio-v3-8b-stt-v2 | low | WER | Unknown | bosonai | 402 | bosonai/higgs-audio-v3-8b | 10.14 |
| Cohere-Speech-Tashkeel-2B | medium | **WER — undiacritized** | cohere-transcribe-arabic-07-2026 | NAMAA-Space | 403 | CohereLabs/cohere-transcribe-arabic-07-2026 | 10.13 |
| higgs-audio-v3-stt-v2 | low | WER (%) | higgs-audio-v3-stt | bosonai | 404 | bosonai/higgs-audio-v3-stt | 10.03 |
| whisper-large-v3-turbo-bulgarian-bulgaria | medium | Wer | whisper-large-v3-turbo | sam8000 | 405 | openai/whisper-large-v3-turbo | 9.97 |
| whisper-large-v3-russian | medium | WER | Unknown | antony66 | 406 | antony66/whisper-large-v3-russian | 9.84 |
| hindi-tts-190M | low | WER ↓ | Unknown | 5ivatej | 407 | — | 9.8 |
| whisper-small-singlish-122k | medium | WER | whisper-small | jensenlwt | 408 | openai/whisper-small | 9.69 |
| moonshine-base-vi-gguf | low | WER (FLEURS vi test) | Unknown | handy-computer | 409 | UsefulSensors/moonshine-base-vi | 9.67 |
| whisper-large-v3-it-cv16 | medium | wer | Unknown | fsicoli | 410 | openai/whisper-large | 9.63 |
| parakeet-tdt-0.6b-portuguese | low | WER | parakeet-tdt-0-6b-v3 | yuriyvnv | 411 | nvidia/parakeet-tdt-0.6b-v3 | 9.62 |
| w2vbert-lingala-waxal | medium | Wer | Unknown | sulaimank | 412 | sulaimank/w2v-bert-lingala-109h | 9.61 |
| Nemotron-3.5-ASR-Streaming-0.6B-CoreML-INT8 | low | WER % | nemotron-3-5-asr-streaming-0-6b | aufklarer | 413 | nvidia/nemotron-3.5-asr-streaming-0.6b | 9.59 |
| Nemotron-3.5-ASR-Streaming-0.6B-CoreML-INT8 | low | WER % | nemotron-3-5-asr-streaming-0-6b | playstonex | 414 | nvidia/nemotron-3.5-asr-streaming-0.6b | 9.59 |
| wav2vec2-large-xlsr-53-german-cv9 | medium | Wer | Unknown | oliverguhr | 415 | — | 9.48 |
| whisper-small-ko | low | Wer | Unknown | SungBeom | 416 | — | 9.48 |
| whisper-accent-medium.en | medium | Wer | Unknown | mavleo96 | 417 | openai/whisper-medium.en | 9.39 |
| whisper-large-v3-spanish | medium | Wer | whisper-large-v3 | Berly00 | 418 | openai/whisper-large-v3 | 9.36 |
| parakeet-ctc-0.6b-Vietnamese | low | AVG WER | Unknown | nvidia | 419 | nvidia/parakeet-ctc-0.6b-Vietnamese | 9.3 |
| parakeet_realtime_eou_120m-v1 | medium | WER (%) | Unknown | nvidia | 420 | nvidia/parakeet_realtime_eou_120m-v1 | 9.3 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER nl | voxtral-mini-3b-2507 | juanginer | 421 | mistralai/Voxtral-Mini-3B-2507 | 9.24 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER nl | voxtral-mini-3b-2507 | juanginer | 422 | mistralai/Voxtral-Mini-3B-2507 | 9.24 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER nl | voxtral-mini-3b-2507 | juanginer | 423 | mistralai/Voxtral-Mini-3B-2507 | 9.24 |
| wav2vec2-large-slavic-parlaspeech-hr | low | WER | Unknown | classla | 424 | — | 9.21 |
| whisper-small_ro-80mel | medium | WER | whisper-small | IonGrozea | 425 | openai/whisper-small | 9.21 |
| shook-medium-amharic-2k | medium | Wer | Unknown | b1n1yam | 426 | — | 9.11 |
| whisper-medium-jp | medium | Wer | Unknown | vumichien | 427 | — | 9.04 |
| whisper-large-v3-turbo-russian | medium | WER | whisper-large-v3-turbo | coriollon | 428 | openai/whisper-large-v3-turbo | 8.93 |
| whisper-small_ro-80mel | medium | WER (normalized: locase, punctuation stripped) | whisper-small | IonGrozea | 429 | openai/whisper-small | 8.8 |
| xls-r-300m-danish-mc-v2 | medium | Wer | Unknown | MediaCatch | 430 | MediaCatch/xls-r-300m-danish-mc | 8.75 |
| whisper-persian-v4 | medium | Wer | whisper-large-v3 | nezamisafa | 431 | openai/whisper-large-v3 | 8.73 |
| Pathumma-whisper-th-large-v3 | high | CV18 (WER) | whisper-large-v3 | nectec | 432 | openai/whisper-large-v3 | 8.68 |
| wav2vec2-xls-r-300m-cv7-turkish | low | WER | Unknown | mpoyraz | 433 | — | 8.62 |
| Shenava-Koochik-0.9 | low | **WER** | Unknown | Reza2kn | 434 | nvidia/stt_en_fastconformer_hybrid_large_streaming_multi | 8.57 |
| MOSS-Transcribe-Diarize-0.9B-MLX-5bit | low | WER | moss-transcribe-diarize | aufklarer | 435 | OpenMOSS-Team/MOSS-Transcribe-Diarize | 8.55 |
| kid-whisper-small-myst | medium | Wer | Unknown | aadel4 | 436 | — | 8.51 |
| wav2vec2-large-xlsr-53-spanish-with-lm | high | WER | Unknown | patrickvonplaten | 437 | — | 8.44 |
| gigaam-v3-ctc-gguf | low | WER (FLEURS ru) | Unknown | handy-computer | 438 | ai-sage/GigaAM-v3 | 8.42 |
| distil-small.en | low | Short-Form WER ↓ | Unknown | distil-whisper | 439 | distil-whisper/distil-small.en | 8.4 |
| distil-medium.en | low | Short-Form WER ↓ | Unknown | distil-whisper | 440 | — | 8.4 |
| distil-large-v2 | low | Short-Form WER ↓ | distil-large-v3 | distil-whisper | 441 | distil-whisper/distil-large-v3 | 8.4 |
| MOSS-Transcribe-preview-2B | low | WER (%) | moss-transcribe-preview-2b | OpenMOSS-Team | 442 | OpenMOSS-Team/MOSS-Transcribe-preview-2B | 8.37 |
| parakeet-realtime-eou-120m-coreml | low | WER (test-clean) | Unknown | FluidInference | 443 | nvidia/parakeet_realtime_eou_120m-v1 | 8.29 |
| wav2vec2-xlsr-300m-finnish-lm | high | WER (with LM) | Unknown | Finnish-NLP | 444 | — | 8.16 |
| whisper-large-v3-turbo-fr-quebecois | low | Wer | whisper-large-v3-turbo | ele-sage | 445 | openai/whisper-large-v3-turbo | 8.1 |
| gigaam-v3-rnnt-gguf | low | WER (FLEURS ru) | Unknown | handy-computer | 446 | ai-sage/GigaAM-v3 | 8.08 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER fr | voxtral-mini-3b-2507 | juanginer | 447 | mistralai/Voxtral-Mini-3B-2507 | 8.01 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER fr | voxtral-mini-3b-2507 | juanginer | 448 | mistralai/Voxtral-Mini-3B-2507 | 8.01 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER fr | voxtral-mini-3b-2507 | juanginer | 449 | mistralai/Voxtral-Mini-3B-2507 | 8.01 |
| Freya-TTS | medium | WER | Unknown | freyavoice | 450 | — | 8 |
| nemotron-3.5-asr-streaming-0.6b-gguf | low | WER (FLEURS test en (en-US), offline att_context_size=[56, 13]) | nemotron-3-5-asr-streaming-0-6b | handy-computer | 451 | nvidia/nemotron-3.5-asr-streaming-0.6b | 7.97 |
| nemotron-3.5-asr-streaming-0.6b-gguf | low | WER (FLEURS test en (en-US), offline att_context_size=[56, 13]) | nemotron-3-5-asr-streaming-0-6b | memoravox | 452 | nvidia/nemotron-3.5-asr-streaming-0.6b | 7.97 |
| xls-r-300m-et | low | WER | Unknown | TalTechNLP | 453 | — | 7.9 |
| wav2vec2-base-word-by-word-quran-asr | medium | Wer | Unknown | HamzaSidhu786 | 454 | facebook/wav2vec2-base | 7.9 |
| whisper-large-v3-czech-cv13 | medium | Wer | whisper-large-v3 | mikr | 455 | openai/whisper-large-v3 | 7.89 |
| Nos_ASR-wav2vec2-xls-r-300m-gl | low | WER | Unknown | proxectonos | 456 | facebook/wav2vec2-xls-r-300m | 7.85 |
| ATC-whisper-turbo-v1 | low | WER | whisper-large-v3-turbo | SingularityUS | 457 | openai/whisper-large-v3-turbo | 7.83 |
| whisper-large-v3-turbo-verbatim-1 | medium | Wer | whisper-large-v3-turbo | JacobLinCool | 458 | openai/whisper-large-v3-turbo | 7.73 |
| Qwen3-ASR-1.7B-fp8 | high | WER | qwen3-asr-1-7b | vrfai | 459 | Qwen/Qwen3-ASR-1.7B | 7.6 |
| whisper-large-v3-pt-cv19-fleurs | medium | Wer | whisper-large-v3 | fsicoli | 460 | openai/whisper-large-v3 | 7.56 |
| whisper-tiny-gguf | low | WER (LibriSpeech test-clean) | whisper-tiny | handy-computer | 461 | openai/whisper-tiny | 7.54 |
| LFM2.5-Audio-1.5B | high | Open | lfm2-1-2b | LiquidAI | 462 | LiquidAI/LFM2-1.2B | 7.53 |
| LFM2.5-Audio-1.5B-GGUF | high | Open | lfm2-1-2b | Mungert | 463 | LiquidAI/LFM2-1.2B | 7.53 |
| Shenava-Koochik-v1.0 | high | visualears-golden-6669 — WER | Unknown | Reza2kn | 464 | nvidia/stt_fa_fastconformer_hybrid_large | 7.49 |
| whisper-hu-large-v3-turbo-finetuned | medium | Wer | whisper-large-v3-turbo | sarpba | 465 | openai/whisper-large-v3-turbo | 7.48 |
| whisper-th-medium-combined | medium | WER | whisper-medium | biodatlab | 466 | openai/whisper-medium | 7.42 |
| distil-large-v3.5 | low | Short-Form OOD WER | distil-large-v3-5 | distil-whisper | 467 | distil-whisper/distil-large-v3.5 | 7.3 |
| wav2vec2-2-bart-base | medium | Wer | Unknown | patrickvonplaten | 468 | — | 7.28 |
| whisper-large-v2-pl-v2 | medium | Wer | Unknown | bardsai | 469 | — | 7.28 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER pt | voxtral-mini-3b-2507 | juanginer | 470 | mistralai/Voxtral-Mini-3B-2507 | 7.16 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER pt | voxtral-mini-3b-2507 | juanginer | 471 | mistralai/Voxtral-Mini-3B-2507 | 7.16 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER pt | voxtral-mini-3b-2507 | juanginer | 472 | mistralai/Voxtral-Mini-3B-2507 | 7.16 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER avg-6 | voxtral-mini-3b-2507 | juanginer | 473 | mistralai/Voxtral-Mini-3B-2507 | 7.13 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER avg-6 | voxtral-mini-3b-2507 | juanginer | 474 | mistralai/Voxtral-Mini-3B-2507 | 7.13 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER avg-6 | voxtral-mini-3b-2507 | juanginer | 475 | mistralai/Voxtral-Mini-3B-2507 | 7.13 |
| g-omnivoice | low | WER ↓ | omnivoice | g-group-ai-lab | 476 | k2-fsa/OmniVoice | 7.12 |
| whisper-tiny-ar-quran | medium | Wer | Unknown | tarteel-ai | 477 | — | 7.05 |
| nb-wav2vec2-300m-bokmaal | medium | WER | Unknown | NbAiLab | 478 | — | 7.03 |
| VoiceTut-TTS | medium | WER — English | omnivoice | mohammedaly22 | 479 | k2-fsa/OmniVoice | 7 |
| nemotron-speech-streaming-en-0.6b | medium | **WER (%)** | nemotron-speech-streaming-en-0-6b | nvidia | 480 | nvidia/nemotron-speech-streaming-en-0.6b | 6.93 |
| nemotron-speech-streaming-en-0.6b-gguf | medium | **WER (%)** | nemotron-speech-streaming-en-0-6b | handy-computer | 481 | nvidia/nemotron-speech-streaming-en-0.6b | 6.93 |
| nemotron-3.5-asr-streaming-GGUF | high | Published WER | nemotron-3-5-asr-streaming-0-6b | cstr | 482 | nvidia/nemotron-3.5-asr-streaming-0.6b | 6.93 |
| nemotron-speech-streaming-en-0.6b-gguf | medium | **WER (%)** | nemotron-speech-streaming-en-0-6b | memoravox | 483 | nvidia/nemotron-speech-streaming-en-0.6b | 6.93 |
| wav2vec2-large-xlsr-catala | low | WER | Unknown | softcatala | 484 | — | 6.92 |
| moonshine-base-de-fidoriel-GGUF | medium | WER | moonshine-base | cstr | 485 | fidoriel/moonshine-base-de | 6.9 |
| moonshine-base-de | medium | WER | moonshine-base | fidoriel | 486 | UsefulSensors/moonshine-base | 6.9 |
| wavlm-libri-clean-100h-base-plus | medium | Wer | Unknown | patrickvonplaten | 487 | — | 6.83 |
| asr-conformer-loquacious | low | validation WER | Unknown | speechbrain | 488 | — | 6.8 |
| wavlm-libri-clean-100h-base | medium | Wer | Unknown | patrickvonplaten | 489 | — | 6.75 |
| Qwen3-ASR-1.7B-NL | low | Zero-shot WER | CER (zero-shot → fine-tuned) | qwen3-asr-1-7b | yuriyvnv | 490 | Qwen/Qwen3-ASR-1.7B | 6.7 |
| f5-tts-quebec-french-finetune | medium | WER ↓ (faster-whisper large-v3) | Unknown | tontate | 491 | RASPIAUDIO/F5-French-MixedSpeakers-reduced | 6.7 |
| wav2vec2-large-danish-npsc-nst | medium | Wer | Unknown | NbAiLab | 492 | chcaa/xls-r-300m-danish | 6.69 |
| CrisperWhisper | medium | **Average WER** | whisper-large-v3 | nyralabs | 493 | openai/whisper-large-v3 | 6.66 |
| CrisperWhisper | medium | **Average WER** | whisper-large-v3 | unsloth | 494 | openai/whisper-large-v3 | 6.66 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER en (no MLS) | voxtral-mini-3b-2507 | juanginer | 495 | mistralai/Voxtral-Mini-3B-2507 | 6.65 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER en (no MLS) | voxtral-mini-3b-2507 | juanginer | 496 | mistralai/Voxtral-Mini-3B-2507 | 6.65 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER en (no MLS) | voxtral-mini-3b-2507 | juanginer | 497 | mistralai/Voxtral-Mini-3B-2507 | 6.65 |
| whisper-medium_ro-80mel | medium | WER | whisper-medium | IonGrozea | 498 | openai/whisper-medium | 6.63 |
| whisper-medium-portuguese | medium | Wer | Unknown | pierreguillou | 499 | — | 6.6 |
| w2v-bert-uk | medium | WER | w2v-bert-2-0 | Yehor | 500 | facebook/w2v-bert-2.0 | 6.6 |
| whisper-th-large-v3-combined | medium | WER | whisper-large-v3 | biodatlab | 501 | openai/whisper-large-v3 | 6.59 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER it | voxtral-mini-3b-2507 | juanginer | 502 | mistralai/Voxtral-Mini-3B-2507 | 6.5 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER it | voxtral-mini-3b-2507 | juanginer | 503 | mistralai/Voxtral-Mini-3B-2507 | 6.5 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER it | voxtral-mini-3b-2507 | juanginer | 504 | mistralai/Voxtral-Mini-3B-2507 | 6.5 |
| Fasih-TTS-V1 | low | WER · Whisper | Unknown | NightPrince | 505 | coqui/XTTS-v2 | 6.5 |
| parakeet-tdt-0.6b-v3 | low | **Avg WER** | parakeet-tdt-0-6b-v3 | nvidia | 506 | nvidia/parakeet-tdt-0.6b-v3 | 6.34 |
| canary-1b-v2-GGUF | medium | Open ASR WER (avg) | Unknown | cstr | 507 | nvidia/canary-1b-v2 | 6.34 |
| nb-wav2vec2-1b-bokmaal | medium | WER | Unknown | NbAiLab | 508 | — | 6.33 |
| Borealis | low | Средний WER | Unknown | Vikhrmodels | 509 | — | 6.33 |
| Qwen3-ASR-0.6B-hf | high | Mean WER | qwen3-asr-0-6b-hf | Qwen | 510 | Qwen/Qwen3-ASR-0.6B-hf | 6.31 |
| Qwen3-ASR-0.6B-hf | high | Mean WER | qwen3-asr-0-6b | bezzam | 511 | Qwen/Qwen3-ASR-0.6B | 6.31 |
| whisper-medium_ro-80mel | medium | WER (normalized: locase, punctuation stripped) | whisper-medium | IonGrozea | 512 | openai/whisper-medium | 6.3 |
| whisper-large-id | medium | Wer | Unknown | cahya | 513 | — | 6.25 |
| higgs-audio-v3-8b-stt | low | WER | Unknown | bosonai | 514 | bosonai/higgs-audio-v3-8b | 6.23 |
| Phi-4-multimodal-instruct | medium | Word Error Rate (WER) | phi-4-multimodal-instruct | microsoft | 515 | microsoft/Phi-4-multimodal-instruct | 6.14 |
| Phi-4-multimodal-instruct | medium | Word Error Rate (WER) | phi-4-multimodal-instruct | Lexius | 516 | microsoft/Phi-4-multimodal-instruct | 6.14 |
| whisper-small-ckb | medium | Wer | whisper-small | razhan | 517 | openai/whisper-small | 6.13 |
| thewhisper-large-v3-turbo | medium | **Mean WER** | whisper-large-v3-turbo | TheStageAI | 518 | openai/whisper-large-v3-turbo | 6.12 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER de | voxtral-mini-3b-2507 | juanginer | 519 | mistralai/Voxtral-Mini-3B-2507 | 6.06 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER de | voxtral-mini-3b-2507 | juanginer | 520 | mistralai/Voxtral-Mini-3B-2507 | 6.06 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER de | voxtral-mini-3b-2507 | juanginer | 521 | mistralai/Voxtral-Mini-3B-2507 | 6.06 |
| whisper-small-id | medium | Wer | Unknown | cahya | 522 | — | 6.06 |
| parakeet-tdt-0.6b-v2 | high | **Avg WER** | Unknown | nvidia | 523 | nvidia/parakeet-tdt-0.6b-v2 | 6.05 |
| parakeet-tdt-0.6b-v2-gguf | high | **Avg WER** | Unknown | handy-computer | 524 | nvidia/parakeet-tdt-0.6b-v2 | 6.05 |
| wav2vec2-large-100k-voxpopuli-catala | low | WER | Unknown | ccoreilly | 525 | — | 5.98 |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 | medium | HF-ASR (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | nvidia | 526 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 5.95 |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-FP8 | medium | HF-ASR (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | nvidia | 527 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 5.95 |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | medium | HF-ASR (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | nvidia | 528 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 5.95 |
| NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF | medium | HF-ASR (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | unsloth | 529 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 5.95 |
| NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning | medium | HF-ASR (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | unsloth | 530 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 5.95 |
| moonshine_tiny_pt_v08 | medium | Wer | Unknown | aomocelin | 531 | aomocelin/moonshine_tiny_pt_v05 | 5.86 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER es | voxtral-mini-3b-2507 | juanginer | 532 | mistralai/Voxtral-Mini-3B-2507 | 5.82 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER es | voxtral-mini-3b-2507 | juanginer | 533 | mistralai/Voxtral-Mini-3B-2507 | 5.82 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER es | voxtral-mini-3b-2507 | juanginer | 534 | mistralai/Voxtral-Mini-3B-2507 | 5.82 |
| whisper-tiny.en-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 535 | openai/whisper-tiny.en | 5.77 |
| whisper-base-ar-quran | medium | Wer | whisper-base | tarteel-ai | 536 | openai/whisper-base | 5.75 |
| wav2vec2-xlsr-1b-finnish-lm-v2 | high | WER (with LM) | Unknown | Finnish-NLP | 537 | — | 5.65 |
| Qwen3-ASR-1.7B-hf | high | Mean WER | qwen3-asr-1-7b-hf | Qwen | 538 | Qwen/Qwen3-ASR-1.7B-hf | 5.59 |
| Qwen3-ASR-1.7B-hf | high | Mean WER | qwen3-asr-1-7b | bezzam | 539 | Qwen/Qwen3-ASR-1.7B | 5.59 |
| wav2vec2-large-slavic-parlaspeech-hr-lm | low | WER | Unknown | classla | 540 | — | 5.56 |
| Zipformer-cr-ctc-transducer-XL-290M | medium | Average WER | Unknown | soundsgoodai | 541 | — | 5.56 |
| Whisper-Large-v3-Turbo-CoreML | low | WER | whisper-large-v3-turbo | aufklarer | 542 | openai/whisper-large-v3-turbo | 5.55 |
| granite-4.0-1b-speech | high | **Average WER** | granite-4-0-1b-base | ibm-granite | 543 | ibm-granite/granite-4.0-1b-base | 5.52 |
| granite-4.0-1b-speech-ONNX | low | **Average WER** | granite-4-0-1b-speech | onnx-community | 544 | ibm-granite/granite-4.0-1b-speech | 5.52 |
| gigaam-v3-e2e-ctc-gguf | low | WER (FLEURS ru) | Unknown | handy-computer | 545 | ai-sage/GigaAM-v3 | 5.5 |
| gigaam-v3-e2e-ctc-gguf | low | WER (FLEURS ru) | Unknown | memoravox | 546 | ai-sage/GigaAM-v3 | 5.5 |
| cohere-transcribe-03-2026-GGUF | medium | WER | cohere-transcribe-03-2026 | cstr | 547 | CohereLabs/cohere-transcribe-03-2026 | 5.42 |
| fastconformer-quran | high | Held-out Quran benchmark (EveryAyah unseen + QUL unseen + tlog phone, 600 clips) — normalized overall WER | Unknown | Muno459 | 548 | nvidia/stt_ar_fastconformer_hybrid_large_pcd_v1.0 | 5.41 |
| gigaam-v3-e2e-rnnt-gguf | low | WER (FLEURS ru) | Unknown | handy-computer | 549 | ai-sage/GigaAM-v3 | 5.35 |
| gigaam-v3-e2e-rnnt-gguf | low | WER (FLEURS ru) | Unknown | memoravox | 550 | ai-sage/GigaAM-v3 | 5.35 |
| granite-speech-4.1-2b-plus | high | **Average WER** | granite-4-0-1b-base | ibm-granite | 551 | ibm-granite/granite-4.0-1b-base | 5.33 |
| whisper-base-gguf | low | WER (LibriSpeech test-clean) | whisper-base | handy-computer | 552 | openai/whisper-base | 5.1 |
| whisper-base-gguf | low | WER (LibriSpeech test-clean) | whisper-base | memoravox | 553 | openai/whisper-base | 5.1 |
| cohere-transcribe-arabic-07-2026-MLX-Mixed-2bit3bit4bit | low | WER | cohere-transcribe-arabic-07-2026 | MarkChen1214 | 554 | CohereLabs/cohere-transcribe-arabic-07-2026 | 5.09 |
| gigaam-multilingual-mlx | low | Core macro WER | Unknown | ai-babai | 555 | ai-sage/GigaAM-Multilingual | 5.07 |
| gigaam-multilingual-mlx-int4-g64 | low | Core macro WER | Unknown | ai-babai | 556 | ai-babai/gigaam-multilingual-mlx | 5.07 |
| gigaam-multilingual-mlx-int8-g64 | low | Core macro WER | Unknown | ai-babai | 557 | ai-babai/gigaam-multilingual-mlx | 5.07 |
| gigaam-multilingual-mlx-int6-g64 | low | Core macro WER | Unknown | ai-babai | 558 | ai-babai/gigaam-multilingual-mlx | 5.07 |
| best_for_now | medium | Word Error Rate (WER) | whisper-small | MosesJoshuaCoker | 559 | openai/whisper-small | 5.05 |
| wavlm-libri-clean-100h-large | medium | Wer | Unknown | patrickvonplaten | 560 | — | 4.91 |
| whisper-tiny-yoruba-v2 | medium | Wer | whisper-tiny | EYEDOL | 561 | openai/whisper-tiny | 4.91 |
| MOSS-TTSD-v1.0 | high | ZH - WER | Unknown | OpenMOSS-Team | 562 | — | 4.85 |
| whisper-large-v3-lv-late-cv19 | low | WER | whisper-large-v3 | AiLab-IMCS-UL | 563 | openai/whisper-large-v3 | 4.8 |
| wav2vec2-large-xlsr-korean | medium | WER | Unknown | kresnik | 564 | — | 4.74 |
| Voxtral-Mini-3B-2507-AWQ4-ARG | medium | WER es-AR | voxtral-mini-3b-2507 | juanginer | 565 | mistralai/Voxtral-Mini-3B-2507 | 4.69 |
| Voxtral-Mini-3B-2507-AWQ4-GEN | medium | WER es-AR | voxtral-mini-3b-2507 | juanginer | 566 | mistralai/Voxtral-Mini-3B-2507 | 4.69 |
| Voxtral-Mini-3B-2507-AWQ4-EN | medium | WER es-AR | voxtral-mini-3b-2507 | juanginer | 567 | mistralai/Voxtral-Mini-3B-2507 | 4.69 |
| moonshine-tiny-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 568 | UsefulSensors/moonshine-tiny | 4.58 |
| BanglaASR | high | Bangla Mozilla Common Voice validation — WER | Unknown | bangla-speech-processing | 569 | — | 4.58 |
| whisper-medium-hindi-hinglish-asr-fine-tuned | medium | Word Error Rate (WER) | whisper-medium | shujaAK | 570 | openai/whisper-medium | 4.56 |
| whisper-medium-hindi-hinglish-asr-fine-tuned | medium | WER | whisper-medium | shujaAK | 571 | openai/whisper-medium | 4.56 |
| moonshine-tiny-GGUF | medium | WER | Unknown | cstr | 572 | UsefulSensors/moonshine-tiny | 4.55 |
| moonshine-streaming-tiny-gguf | low | WER (LibriSpeech test-clean) | moonshine-streaming-tiny | handy-computer | 573 | UsefulSensors/moonshine-streaming-tiny | 4.53 |
| whisper-large-v3-high-mixed-nl | medium | **Test WER (Common Voice)** | whisper-large-v3 | yuriyvnv | 574 | openai/whisper-large-v3 | 4.43 |
| dots.tts-mf | low | hard-en WER↓ | Unknown | dots-studio | 575 | rednote-hilab/dots.tts-soar | 4.4 |
| qwen3-asr-0.6b-coreml | low | WER | qwen3-asr-0-6b | FluidInference | 576 | Qwen/Qwen3-ASR-0.6B | 4.4 |
| Inflect-Nano-v2 | low | Two-ASR semantic WER ↓ | Unknown | owensong | 577 | owensong/Inflect-Nano-v1 | 4.21 |
| Inflect-Nano-v2-GGUF | low | Two-ASR semantic WER ↓ | Unknown | remixerdec | 578 | owensong/Inflect-Nano-v2 | 4.21 |
| whisper-base.en-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 579 | openai/whisper-base.en | 4.14 |
| wav2vec2-large-uralic-voxpopuli-v2-finnish | high | WER (with LM) | Unknown | Finnish-NLP | 580 | — | 4.13 |
| Inflect-Micro-v2 | low | Two-ASR semantic WER ↓ | Unknown | owensong | 581 | owensong/Inflect-Micro-v2 | 3.99 |
| Inflect-Micro-v2-GGUF | low | Two-ASR semantic WER ↓ | Unknown | remixerdec | 582 | owensong/Inflect-Micro-v2 | 3.99 |
| WhisperLv3-FT | medium | Wer | whisper-large-v3 | inesc-id | 583 | openai/whisper-large-v3 | 3.97 |
| whisper-medium-id | medium | Wer | whisper-medium | cahya | 584 | openai/whisper-medium | 3.83 |
| whisper-large-v3-high-mixed-nl | medium | **Validation WER** | whisper-large-v3 | yuriyvnv | 585 | openai/whisper-large-v3 | 3.57 |
| whisper-large-v3-high-mixed-nl | low | Val WER | whisper-large-v3 | yuriyvnv | 586 | openai/whisper-large-v3 | 3.56 |
| whisper-large-v3-turbo-hr-parla | medium | Wer | whisper-large-v3 | GoranS | 587 | openai/whisper-large-v3 | 3.52 |
| wav2vec2-1b-npsc-nst-bokmaal-repaired | medium | Wer | Unknown | NbAiLab | 588 | facebook/wav2vec2-xls-r-1b | 3.49 |
| whisper-tiny-ar-quran-mix-norm | medium | Wer | whisper-tiny | deepdml | 589 | openai/whisper-tiny | 3.39 |
| whisper-small-gguf | low | WER (LibriSpeech test-clean) | whisper-small | handy-computer | 590 | openai/whisper-small | 3.33 |
| moonshine-base-gguf | low | WER (LibriSpeech test-clean) | moonshine-base | handy-computer | 591 | UsefulSensors/moonshine-base | 3.28 |
| SenseVoiceSmall-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 592 | FunAudioLLM/SenseVoiceSmall | 3.13 |
| whisper-small-finetuned-medical3 | medium | Wer | whisper-small | Johnyquest7 | 593 | openai/whisper-small | 3.12 |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 | medium | Tedium Long (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | nvidia | 594 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 3.11 |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-FP8 | medium | Tedium Long (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | nvidia | 595 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 3.11 |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | medium | Tedium Long (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | nvidia | 596 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 3.11 |
| NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF | medium | Tedium Long (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | unsloth | 597 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 3.11 |
| NVIDIA-Nemotron-3-Nano-Omni-30B-A3B-Reasoning | medium | Tedium Long (WER↓) | nemotron-3-nano-omni-30b-a3b-reasoning-bf16 | unsloth | 598 | nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 3.11 |
| parakeet-tdt-0.6b-v3-coreml | low | WER vs fp16 ref | parakeet-tdt-0-6b-v3 | mweinbach1 | 599 | nvidia/parakeet-tdt-0.6b-v3 | 3.1 |
| whisper-small.en-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 600 | openai/whisper-small.en | 3.09 |
| asr-crdnn-rnnlm-librispeech | low | Test WER | Unknown | speechbrain | 601 | — | 3.09 |
| moonshine_tiny_pt_v06 | medium | Wer | Unknown | aomocelin | 602 | aomocelin/moonshine_tiny_pt_v05 | 3 |
| parakeet-tdt-ctc-110m-coreml | low | WER | Unknown | FluidInference | 603 | nvidia/parakeet-tdt_ctc-110m | 3 |
| w2vbert-shona-waxal | medium | Wer | w2v-bert-2-0 | sulaimank | 604 | sulaimank/W2V2_Bert_Afrivoice_FLEURS_Shona_100hr_v1 | 2.96 |
| mtl-mimic-voicebank | low | Valid WER | Unknown | speechbrain | 605 | — | 2.89 |
| dots.tts-soar | low | Avg WER↓ | Unknown | dots-studio | 606 | rednote-hilab/dots.tts-base | 2.8 |
| dots.tts-base | low | Avg WER↓ | Unknown | dots-studio | 607 | rednote-hilab/dots.tts-base | 2.8 |
| ZONOS2-GGUF | low | WER ↓ | Unknown | Zyphra | 608 | Zyphra/ZONOS2 | 2.79 |
| whisper-medium.en-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 609 | openai/whisper-medium.en | 2.74 |
| whisper-large-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 610 | openai/whisper-large | 2.72 |
| asr-streaming-conformer-librispeech | low | Test clean WER | Unknown | speechbrain | 611 | — | 2.72 |
| Voila-Tokenizer | low | LibriSpeech test-clean (WER) | Unknown | maitrix-org | 612 | — | 2.7 |
| whisper-large-v2-gguf | low | WER (LibriSpeech test-clean) | whisper-large-v2 | handy-computer | 613 | openai/whisper-large-v2 | 2.68 |
| whisper-medium-gguf | low | WER (LibriSpeech test-clean) | whisper-medium | handy-computer | 614 | openai/whisper-medium | 2.63 |
| whisper-medium-gguf | low | WER (LibriSpeech test-clean) | whisper-medium | memoravox | 615 | openai/whisper-medium | 2.63 |
| cohere-transcribe-03-2026-CoreML-6bit | low | WER vs Ground Truth | cohere-transcribe-03-2026 | BarathwajAnandan | 616 | CohereLabs/cohere-transcribe-03-2026 | 2.58 |
| CosyVoice2-0.5B | high | Open-Source | Unknown | FunAudioLLM | 617 | FunAudioLLM/CosyVoice2-0.5B | 2.57 |
| whisper-small-ar-quran-mix-norm | medium | Wer | whisper-small | deepdml | 618 | openai/whisper-small | 2.54 |
| moonshine-streaming-small-gguf | low | WER (LibriSpeech test-clean) | moonshine-streaming-small | handy-computer | 619 | UsefulSensors/moonshine-streaming-small | 2.53 |
| whisperkit-coreml_01-30-24 | low | WER | Unknown | argmaxinc | 620 | — | 2.44 |
| parakeet-tdt_ctc-110m-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 621 | nvidia/parakeet-tdt_ctc-110m | 2.43 |
| VibeVoice-Realtime-0.5B | low | WER (%) ↓ | qwen-qwen2-5-0-5b | microsoft | 622 | Qwen/Qwen2.5-0.5B | 2.4 |
| VibeVoice-Realtime-0.5B-ONNX | low | WER (%) ↓ | qwen-qwen2-5-0-5b | elbruno | 623 | microsoft/VibeVoice-Realtime-0.5B | 2.4 |
| parakeet-unified-en-0.6b-coreml | low | Aggregate WER\* | Unknown | FluidInference | 624 | nvidia/parakeet-tdt-0.6b-v2 | 2.37 |
| nemotron-speech-streaming-en-0.6b-gguf | low | WER (LibriSpeech test-clean, offline) | nemotron-speech-streaming-en-0-6b | handy-computer | 625 | nvidia/nemotron-speech-streaming-en-0.6b | 2.31 |
| nemotron-speech-streaming-en-0.6b-gguf | low | WER (LibriSpeech test-clean, offline) | nemotron-speech-streaming-en-0-6b | memoravox | 626 | nvidia/nemotron-speech-streaming-en-0.6b | 2.31 |
| Breeze-ASR-25-gguf | low | WER (LibriSpeech test-clean) | whisper-large-v2 | handy-computer | 627 | MediaTek-Research/Breeze-ASR-25 | 2.29 |
| asr-transformer-transformerlm-librispeech | low | Test clean WER | Unknown | speechbrain | 628 | — | 2.27 |
| Fun-CosyVoice3-0.5B-2512 | high | Open-Source | Unknown | FunAudioLLM | 629 | FunAudioLLM/Fun-CosyVoice3-0.5B-2512 | 2.24 |
| multitalker-parakeet-streaming-0.6b-v1-gguf | low | WER (LibriSpeech test-clean, offline) | Unknown | handy-computer | 630 | nvidia/multitalker-parakeet-streaming-0.6b-v1 | 2.19 |
| moonshine-streaming-medium-gguf | low | WER (LibriSpeech test-clean) | moonshine-streaming-medium | handy-computer | 631 | UsefulSensors/moonshine-streaming-medium | 2.16 |
| CosyVoice-300M-SFT | low | Open-Source | Unknown | FunAudioLLM | 632 | — | 2.14 |
| CosyVoice-300M | low | Open-Source | Unknown | FunAudioLLM | 633 | FunAudioLLM/CosyVoice-300M | 2.14 |
| CosyVoice-300M-Instruct | low | Open-Source | Unknown | FunAudioLLM | 634 | FunAudioLLM/CosyVoice-300M | 2.14 |
| Raon-OpenTTS-1B | low | WER (%) ↓ | Unknown | KRAFTON | 635 | — | 2.14 |
| Qwen3-ASR-0.6B-gguf | low | WER (LibriSpeech test-clean) | qwen3-asr-0-6b | handy-computer | 636 | Qwen/Qwen3-ASR-0.6B | 2.11 |
| Voxtral-Mini-4B-Realtime-2602-gguf | high | LibriSpeech test-clean (2,620 utterances, BF16 offline) — WER | voxtral-mini-4b-realtime-2602 | handy-computer | 637 | mistralai/Voxtral-Mini-4B-Realtime-2602 | 2.08 |
| moss-transcribe-diarize-gguf | low | WER (LibriSpeech test-clean) | moss-transcribe-diarize | handy-computer | 638 | OpenMOSS-Team/MOSS-Transcribe-Diarize | 2.08 |
| Voxtral-Mini-4B-Realtime-2602-gguf | high | LibriSpeech test-clean (2,620 utterances, BF16 offline) — WER | voxtral-mini-4b-realtime-2602 | memoravox | 639 | mistralai/Voxtral-Mini-4B-Realtime-2602 | 2.08 |
| asr-conformer-transformerlm-librispeech | low | Test clean WER | Unknown | speechbrain | 640 | — | 2.01 |
| whisper-large-v3-turbo-gguf | low | WER (LibriSpeech test-clean) | whisper-large-v3-turbo | handy-computer | 641 | openai/whisper-large-v3-turbo | 2 |
| parakeet-tdt-0.6b-v3-gguf | low | WER (LibriSpeech test-clean) | parakeet-tdt-0-6b-v3 | handy-computer | 642 | nvidia/parakeet-tdt-0.6b-v3 | 1.95 |
| canary-180m-flash-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 643 | nvidia/canary-180m-flash | 1.94 |
| canary-1b-v2-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 644 | nvidia/canary-1b-v2 | 1.92 |
| canary-1b-v2-gguf | low | WER (LibriSpeech test-clean) | Unknown | memoravox | 645 | nvidia/canary-1b-v2 | 1.92 |
| cohere-transcribe-arabic-07-2026-int4 | low | WER vs. full fp16 | cohere-transcribe-arabic-07-2026 | NAMAA-Space | 646 | CohereLabs/cohere-transcribe-arabic-07-2026 | 1.89 |
| Voxtral-Mini-3B-2507-gguf | low | WER (LibriSpeech test-clean) | voxtral-mini-3b-2507 | handy-computer | 647 | mistralai/Voxtral-Mini-3B-2507 | 1.88 |
| parakeet-ctc-0.6b-gguf | low | WER (LibriSpeech test-clean) | parakeet-ctc-0-6b | handy-computer | 648 | nvidia/parakeet-ctc-0.6b | 1.87 |
| parakeet-tdt_ctc-1.1b-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 649 | nvidia/parakeet-tdt_ctc-1.1b | 1.87 |
| parakeet-ctc-1.1b-gguf | low | WER (LibriSpeech test-clean) | parakeet-ctc-1-1b | handy-computer | 650 | nvidia/parakeet-ctc-1.1b | 1.85 |
| whisper-large-v3-gguf | low | WER (LibriSpeech test-clean) | whisper-large-v3 | handy-computer | 651 | openai/whisper-large-v3 | 1.81 |
| whisper-large-v3-gguf | low | WER (LibriSpeech test-clean) | whisper-large-v3 | memoravox | 652 | openai/whisper-large-v3 | 1.81 |
| MOSS-TTS-GGUF | low | EN WER (%) ↓ | Unknown | OpenMOSS-Team | 653 | — | 1.79 |
| MOSS-TTS-GGUF | low | EN WER (%) ↓ | Unknown | niobures | 654 | — | 1.79 |
| Fun-ASR-Nano-2512-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 655 | FunAudioLLM/Fun-ASR-Nano-2512 | 1.78 |
| Fun-ASR-MLT-Nano-2512-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 656 | FunAudioLLM/Fun-ASR-MLT-Nano-2512 | 1.74 |
| MOSS-TTS | low | Open‑source | Unknown | OpenMOSS-Team | 657 | — | 1.69 |
| MOSS-TTS-Local-Transformer | low | Open‑source | Unknown | OpenMOSS-Team | 658 | OpenMOSS-Team/MOSS-TTS-Local-Transformer | 1.69 |
| MOSS-TTS-GGUF | low | Open‑source | Unknown | OpenMOSS-Team | 659 | — | 1.69 |
| MOSS-TTS-GGUF | low | Open‑source | Unknown | niobures | 660 | — | 1.69 |
| canary-qwen-2.5b-gguf | low | WER (LibriSpeech test-clean) | canary-qwen-2-5b | handy-computer | 661 | nvidia/canary-qwen-2.5b | 1.63 |
| Qwen3-ASR-1.7B-gguf | low | WER (LibriSpeech test-clean) | qwen3-asr-1-7b | handy-computer | 662 | Qwen/Qwen3-ASR-1.7B | 1.62 |
| canary-1b-flash-gguf | low | WER (LibriSpeech test-clean) | canary-1b-flash | handy-computer | 663 | nvidia/canary-1b-flash | 1.62 |
| parakeet-rnnt-0.6b-gguf | low | WER (LibriSpeech test-clean) | parakeet-rnnt-0-6b | handy-computer | 664 | nvidia/parakeet-rnnt-0.6b | 1.62 |
| canary-1b-flash-gguf | low | WER (LibriSpeech test-clean) | canary-1b-flash | memoravox | 665 | nvidia/canary-1b-flash | 1.62 |
| parakeet-rnnt-0.6b-gguf | low | WER (LibriSpeech test-clean) | parakeet-rnnt-0-6b | memoravox | 666 | nvidia/parakeet-rnnt-0.6b | 1.62 |
| spear-xlarge-speech-audio-v2 | low | test-clean WER | Unknown | marcoyang | 667 | — | 1.6 |
| parakeet-unified-en-0.6b-gguf | low | WER (LibriSpeech test-clean, offline) | Unknown | handy-computer | 668 | nvidia/parakeet-unified-en-0.6b | 1.59 |
| parakeet-unified-en-0.6b-gguf | low | WER (LibriSpeech test-clean, offline) | Unknown | memoravox | 669 | nvidia/parakeet-unified-en-0.6b | 1.59 |
| Voxtral-Small-24B-2507-gguf | low | WER (LibriSpeech test-clean) | mistral-small-24b-base-2501 | handy-computer | 670 | mistralai/Voxtral-Small-24B-2507 | 1.56 |
| canary-1b-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 671 | nvidia/canary-1b | 1.55 |
| gepard-1.0 | low | WER ↓ | qwen3-5-full-attn-only-14 | nineninesix | 672 | nineninesix/qwen3_5-full-attn-only-14 | 1.5 |
| granite-speech-4.1-2b-plus-gguf | high | LibriSpeech test-clean (2,620 utterances, BF16 transcribe.cpp) — WER | granite-speech-4-1-2b-plus | handy-computer | 673 | ibm-granite/granite-speech-4.1-2b-plus | 1.49 |
| granite-speech-4.1-2b-plus-gguf | high | LibriSpeech test-clean (2,620 utterances, BF16 transcribe.cpp) — WER | granite-speech-4-1-2b-plus | memoravox | 674 | ibm-granite/granite-speech-4.1-2b-plus | 1.49 |
| parakeet-rnnt-1.1b-gguf | low | WER (LibriSpeech test-clean) | parakeet-rnnt-1-1b | handy-computer | 675 | nvidia/parakeet-rnnt-1.1b | 1.45 |
| inflect-v2-coreml | low | WER* | Unknown | FluidInference | 676 | owensong/Inflect-Micro-v2 | 1.43 |
| granite-4.0-1b-speech-gguf | low | WER (LibriSpeech test-clean) | granite-4-0-1b-speech | handy-computer | 677 | ibm-granite/granite-4.0-1b-speech | 1.42 |
| parakeet-tdt-1.1b-gguf | low | WER (LibriSpeech test-clean) | Unknown | handy-computer | 678 | nvidia/parakeet-tdt-1.1b | 1.39 |
| granite-speech-4.1-2b-gguf | low | WER (LibriSpeech test-clean) | granite-speech-4-1-2b | handy-computer | 679 | ibm-granite/granite-speech-4.1-2b | 1.31 |
| granite-speech-4.1-2b-gguf | low | WER (LibriSpeech test-clean) | granite-speech-4-1-2b | memoravox | 680 | ibm-granite/granite-speech-4.1-2b | 1.31 |
| asr-whisper-large-v2-salt | low | WER | whisper-large-v2 | Sunbird | 681 | openai/whisper-large-v2 | 1.3 |
| granite-speech-4.1-2b-nar-gguf | low | WER (LibriSpeech test-clean) | granite-speech-4-1-2b-nar | handy-computer | 682 | ibm-granite/granite-speech-4.1-2b-nar | 1.29 |
| granite-speech-4.1-2b-nar-gguf | low | WER (LibriSpeech test-clean) | granite-speech-4-1-2b-nar | memoravox | 683 | ibm-granite/granite-speech-4.1-2b-nar | 1.29 |
| cohere-transcribe-03-2026-gguf | low | WER (LibriSpeech test-clean) | cohere-transcribe-03-2026 | handy-computer | 684 | CohereLabs/cohere-transcribe-03-2026 | 1.26 |
| cohere-transcribe-03-2026-gguf | low | WER (LibriSpeech test-clean) | cohere-transcribe-03-2026 | memoravox | 685 | CohereLabs/cohere-transcribe-03-2026 | 1.26 |
| wav2vec2-xls-r-1b-japanese-hiragana-katakana | medium | Wer | Unknown | AndrewMcDowell | 686 | — | 1.01 |
| wav2vec2-vietnamese-asr-v1 | medium | Wer | Unknown | cheesethuong | 687 | nguyenvulebinh/wav2vec2-base-vietnamese-250h | 1 |
| fastconformer-quran-ar | low | Val WER | Unknown | mohammed | 688 | — | 0.38 |
Provenance & metadata (1)
source
- kind
- leaderboard-scrape
- paths
- benchmarks/wer.html, docs/notes/hf-benchmark-contexts.json