| 1 | TrorYongASR-small | — | KrorngAI | 100.0 | high | |
| 2 | TrorYongASR-tiny | — | KrorngAI | 100.0 | high | |
| 3 | SE-DiCoW | — | BUT-FIT | 17.9 | medium | **LibriSpeechMix (2)** |
| 4 | DiCoW_v3_3 | openai/whisper-large-v3-turbo | BUT-FIT | 17.9 | medium | **LibriSpeechMix (2)** |
| 5 | moonshine-streaming-tiny | UsefulSensors/moonshine-streaming-tiny | UsefulSensors | 12.09 | medium | LibriSpeech (other) |
| 6 | moonshine-streaming-medium-gguf | UsefulSensors/moonshine-streaming-medium | handy-computer | 12.09 | medium | LibriSpeech (other) |
| 7 | moonshine-streaming-small-gguf | UsefulSensors/moonshine-streaming-small | handy-computer | 12.09 | medium | LibriSpeech (other) |
| 8 | moonshine-streaming-medium | UsefulSensors/moonshine-streaming-medium | UsefulSensors | 12.09 | medium | LibriSpeech (other) |
| 9 | moonshine-streaming-small | UsefulSensors/moonshine-streaming-small | UsefulSensors | 12.09 | medium | LibriSpeech (other) |
| 10 | moonshine-streaming-tiny-gguf | UsefulSensors/moonshine-streaming-tiny | handy-computer | 12.09 | medium | LibriSpeech (other) |
| 11 | resilientai-team-mcte-mistral-voxtral-realtime-gptq-marlin | mistralai/Voxtral-Mini-4B-Realtime-2602 | JaynadTodkarAmol | 10.0 | medium | LibriSpeech (`openslr/librispeech_asr`) |
| 12 | whisper-podlodka-turbo | openai/whisper-large-v3-turbo | bond005 | 9.76 | medium | rulibrispeech |
| 13 | whisper-large-v3-turbo-russian-codeswitch | coriollon/whisper-large-v3-turbo-russian | coriollon | 8.5 | medium | RuLibriSpeech |
| 14 | Qwen3-ASR-1.7B-Swift | mlx-community/Qwen3-ASR-1.7B-8bit | YoozLabs | 6.3 | medium | LibriSpeech-style English |
| 15 | VibeVoice-ASR-HF | microsoft/VibeVoice-ASR-HF | microsoft | 5.51 | medium | librispeech_test.other |
| 16 | parakeet-tdt-ctc-110m-coreml | nvidia/parakeet-tdt_ctc-110m | FluidInference | 5.2 | medium | LibriSpeech test-other |
| 17 | Nemotron-Speech-Streaming-0.6B-CoreML-INT8 | nvidia/nemotron-speech-streaming-en-0.6b | aufklarer | 4.84 | medium | LibriSpeech test-other |
| 18 | Voila-Tokenizer | — | maitrix-org | 4.8 | medium | **Voila (w/o LibriSpeech train split)** |
| 19 | asr-conformer-loquacious | — | speechbrain | 4.6 | low | |
| 20 | moonshine-streaming-tiny | UsefulSensors/moonshine-streaming-tiny | UsefulSensors | 4.49 | medium | LibriSpeech (clean) |
| 21 | moonshine-streaming-medium-gguf | UsefulSensors/moonshine-streaming-medium | handy-computer | 4.49 | medium | LibriSpeech (clean) |
| 22 | moonshine-streaming-small-gguf | UsefulSensors/moonshine-streaming-small | handy-computer | 4.49 | medium | LibriSpeech (clean) |
| 23 | moonshine-streaming-medium | UsefulSensors/moonshine-streaming-medium | UsefulSensors | 4.49 | medium | LibriSpeech (clean) |
| 24 | moonshine-streaming-small | UsefulSensors/moonshine-streaming-small | UsefulSensors | 4.49 | medium | LibriSpeech (clean) |
| 25 | moonshine-streaming-tiny-gguf | UsefulSensors/moonshine-streaming-tiny | handy-computer | 4.49 | medium | LibriSpeech (clean) |
| 26 | qwen3-asr-0.6b-coreml | Qwen/Qwen3-ASR-0.6B | FluidInference | 4.4 | medium | LibriSpeech test-clean (2620 files) |
| 27 | Fun-ASR-MLT-Nano-2512-gguf | FunAudioLLM/Fun-ASR-MLT-Nano-2512 | handy-computer | 4.19 | medium | Librispeech-other |
| 28 | Fun-ASR-Nano-2512-gguf | FunAudioLLM/Fun-ASR-Nano-2512 | handy-computer | 4.19 | medium | Librispeech-other |
| 29 | Fun-ASR-Nano-2512 | FunAudioLLM/Fun-ASR-Nano-2512 | FunAudioLLM | 4.19 | medium | Librispeech-other |
| 30 | Fun-ASR-MLT-Nano-2512 | FunAudioLLM/Fun-ASR-MLT-Nano-2512 | FunAudioLLM | 4.19 | medium | Librispeech-other |
| 31 | CrisperWhisper | openai/whisper-large-v3 | nyralabs | 3.97 | medium | [LibriSpeech other](https://huggingface.co/datasets/openslr/librispeech_asr) |
| 32 | CrisperWhisper | openai/whisper-large-v3 | unsloth | 3.97 | medium | [LibriSpeech other](https://huggingface.co/datasets/openslr/librispeech_asr) |
| 33 | thewhisper-large-v3-turbo | openai/whisper-large-v3-turbo | TheStageAI | 3.77 | medium | **LibriSpeech Other** |
| 34 | Qwen2.5-Omni-7B-AWQ | Qwen/Qwen2.5-Omni-7B | Qwen | 3.4 | medium | LibriSpeech test-other |
| 35 | Qwen2.5-Omni-7B-GPTQ-Int4 | Qwen/Qwen2.5-Omni-7B | Qwen | 3.4 | medium | LibriSpeech test-other |
| 36 | whisper-large-v3-turbo-german | primeline/whisper-large-v3-german | primeline | 3.2 | medium | multilingual librispeech |
| 37 | whisper-large-v3-turbo-german-ct2 | primeline/whisper-large-v3-german | jimmymeister | 3.2 | medium | multilingual librispeech |
| 38 | Zipformer-cr-ctc-transducer-XL-290M | — | soundsgoodai | 3.01 | medium | LibriSpeech Other |
| 39 | parakeet-tdt-ctc-110m-coreml | nvidia/parakeet-tdt_ctc-110m | FluidInference | 3.0 | medium | LibriSpeech test-clean |
| 40 | MOSS-Transcribe-preview-2B | OpenMOSS-Team/MOSS-Transcribe-preview-2B | OpenMOSS-Team | 2.84 | medium | LibriSpeech test.other |
| 41 | higgs-audio-v3-stt-v2 | bosonai/higgs-audio-v3-stt | bosonai | 2.8 | medium | LibriSpeech other |
| 42 | granite-speech-4.1-2b-nar-gguf | ibm-granite/granite-speech-4.1-2b-nar | handy-computer | 2.75 | medium | LibriSpeech other |
| 43 | granite-speech-4.1-2b-nar-gguf | ibm-granite/granite-speech-4.1-2b-nar | memoravox | 2.75 | medium | LibriSpeech other |
| 44 | Voila-Tokenizer | — | maitrix-org | 2.7 | medium | **Voila (with LibriSpeech train split)** |
| 45 | parakeet-primeline | — | primeline | 2.6 | low | |
| 46 | cohere-transcribe-arabic-07-2026-MLX-Mixed-2bit3bit4bit | CohereLabs/cohere-transcribe-arabic-07-2026 | MarkChen1214 | 2.55 | medium | LibriSpeech test-other |
| 47 | kotoba-whisper-bilingual-v1.0 | — | kotoba-tech | 2.4 | low | |
| 48 | higgs-audio-v3-8b-stt-v2 | bosonai/higgs-audio-v3-8b | bosonai | 2.38 | medium | LibriSpeech Other |
| 49 | higgs-audio-v3-8b-stt | bosonai/higgs-audio-v3-8b | bosonai | 2.34 | medium | LibriSpeech Other |
| 50 | Nemotron-Speech-Streaming-0.6B-CoreML-INT8 | nvidia/nemotron-speech-streaming-en-0.6b | aufklarer | 2.32 | medium | LibriSpeech test-clean |
| 51 | VibeVoice-ASR-HF | microsoft/VibeVoice-ASR-HF | microsoft | 2.2 | medium | librispeech_test.clean |
| 52 | LFM2-Audio-1.5B | LiquidAI/LFM2-1.2B | LiquidAI | 2.01 | high | Open |
| 53 | Fun-ASR-MLT-Nano-2512-gguf | FunAudioLLM/Fun-ASR-MLT-Nano-2512 | handy-computer | 2.0 | medium | Librispeech-clean |
| 54 | Fun-ASR-Nano-2512-gguf | FunAudioLLM/Fun-ASR-Nano-2512 | handy-computer | 2.0 | medium | Librispeech-clean |
| 55 | Fun-ASR-Nano-2512 | FunAudioLLM/Fun-ASR-Nano-2512 | FunAudioLLM | 2.0 | medium | Librispeech-clean |
| 56 | Fun-ASR-MLT-Nano-2512 | FunAudioLLM/Fun-ASR-MLT-Nano-2512 | FunAudioLLM | 2.0 | medium | Librispeech-clean |
| 57 | LFM2.5-Audio-1.5B | LiquidAI/LFM2-1.2B | LiquidAI | 1.95 | high | Open |
| 58 | LFM2.5-Audio-1.5B-GGUF | LiquidAI/LFM2-1.2B | Mungert | 1.95 | high | Open |
| 59 | FireRedASR-AED-L | — | FireRedTeam | 1.93 | high | |
| 60 | thewhisper-large-v3-turbo | openai/whisper-large-v3-turbo | TheStageAI | 1.83 | medium | **LibriSpeech Clean** |
| 61 | CrisperWhisper | openai/whisper-large-v3 | nyralabs | 1.74 | medium | [LibriSpeech clean](https://huggingface.co/datasets/openslr/librispeech_asr) |
| 62 | stt_en_fastconformer_hybrid_large_pc | nvidia/stt_en_fastconformer_hybrid_large_pc | nvidia | 1.74 | low | |
| 63 | CrisperWhisper | openai/whisper-large-v3 | unsloth | 1.74 | medium | [LibriSpeech clean](https://huggingface.co/datasets/openslr/librispeech_asr) |
| 64 | Ming-Lite-Omni | inclusionAI/Ling-lite | inclusionAI | 1.44 | high | |
| 65 | higgs-audio-v3-stt-v2 | bosonai/higgs-audio-v3-stt | bosonai | 1.39 | medium | LibriSpeech clean |
| 66 | Zipformer-cr-ctc-transducer-XL-290M | — | soundsgoodai | 1.31 | medium | LibriSpeech Clean |
| 67 | granite-speech-4.1-2b-nar-gguf | ibm-granite/granite-speech-4.1-2b-nar | handy-computer | 1.29 | medium | LibriSpeech clean |
| 68 | granite-speech-4.1-2b-nar-gguf | ibm-granite/granite-speech-4.1-2b-nar | memoravox | 1.29 | medium | LibriSpeech clean |
| 69 | higgs-audio-v3-8b-stt-v2 | bosonai/higgs-audio-v3-8b | bosonai | 1.25 | medium | LibriSpeech Clean |
| 70 | Ming-Lite-Omni-1.5 | inclusionAI/Ming-Lite-Omni | inclusionAI | 1.25 | high | Average on All/Open-source Benchmarks(↓) |
| 71 | higgs-audio-v3-8b-stt | bosonai/higgs-audio-v3-8b | bosonai | 1.24 | medium | LibriSpeech Clean |
| 72 | MOSS-Transcribe-preview-2B | OpenMOSS-Team/MOSS-Transcribe-preview-2B | OpenMOSS-Team | 1.21 | medium | LibriSpeech test.clean |
| 73 | cohere-transcribe-arabic-07-2026-MLX-Mixed-2bit3bit4bit | CohereLabs/cohere-transcribe-arabic-07-2026 | MarkChen1214 | 1.14 | medium | LibriSpeech test-clean |