All tasks / Models
Higgs Audio v3 STT
Boson AI's open-weights speech recognition model from the Higgs Audio v3 family.
- Boson AI
- Open weights
- Released 25 Mar 2026
Where it ranks
- ATranscriptionModels · Local · 5.9% on Hugging Face Open ASR Leaderboard (English, short-form average WER) (secondary benchmark)
Facts
- Vendor
- Boson AI
- Family
- Higgs Audio STT
- Released
- 25 Mar 2026
- Weights
- Open: download and run it yourself
- License
- Apache-2.0
- Parameters
- 2.68B
- GPU memory to run locally
- ~6 GB (Derived: about 2.68B parameters is roughly 5.4 GB of bf16 weights plus runtime overhead; not an official requirement.)
- Access
- Free (open weights)
- Last verified
- 19 Oct 2026
Benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| Hugging Face Open ASR Leaderboard (English, short-form average WER) | 5.9% | bosonai/higgs-audio-v3-sttAverage WER over the board's default datasets (English short-form, version 17-10-2026); rank 15 of 56; read from the Space's /config table data. | 17 Oct 2026 | huggingface.co |
Where you can use it
No tracked tool lists this model yet; as open weights it runs in local runtimes.