HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks
Home/Orgs/bosonai/bosonai/higgs-audio-v3-stt

bosonai/higgs-audio-v3-stt

family misc · 4 variants · unknown

provider
bosonai
released
—
huggingface
https://huggingface.co/bosonai/higgs-audio-v3-stt
best scores
WER: 10.03WER AMI: 10.03WER Earnings-22: 8.95WER GigaSpeech: 8.16WER LibriSpeech: 2.8WER SPGISpeech: 3.76WER TED-LIUM: 2.76WER VoxPopuli: 6.07

Best scores per benchmark

benchmarkcategorybest score
WERspeech-asr10.03
WER AMIspeech-asr10.03
WER Earnings-22speech-asr8.95
WER GigaSpeechspeech-asr8.16
WER VoxPopulispeech-asr6.07
WER SPGISpeechspeech-asr3.76
WER LibriSpeechspeech-asr2.8
WER TED-LIUMspeech-asr2.76

Variants (4)

variantkindquantbitsvRAMdlorgyear
higgs-audio-v3-sttoriginalbf1616.07,105bosonai2026
higgs-audio-v3-stt-v2originalbf1616.0931bosonai2026
higgs-audio-v3-stt-GGUFggufgguf16.0778cstr2026
higgs-audio-v3-sttoriginalbf1616.017percebeth2026
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.