All tasks / Models
Higgs TTS 3 (4B)
Boson AI's open-weights 4B conversational text-to-speech model with zero-shot voice cloning in 100+ languages.
- Boson AI
- Open weights
- Released 4 Jun 2026
Where it ranks
- CText to speechModels · Local · 1,042 on Artificial Analysis Speech Arena: Text to Speech (provider voices)
Facts
- Vendor
- Boson AI
- Family
- Higgs TTS
- Released
- 4 Jun 2026
- Weights
- Open: download and run it yourself
- License
- Boson Higgs TTS 3 Research and Non-Commercial License
- Parameters
- 4.65B
- GPU memory to run locally
- ~12 GB (Derived: 4.65B parameters (Hugging Face safetensors count) x 2 bytes = ~9.3 GB of bf16 weights, plus overhead.)
- API price
- $15 per 1M characters (Artificial Analysis normalised price for Higgs Audio V3 TTS (USD per 1M characters).)
- Access
- Free (open weights)
- Last verified
- 18 Oct 2026
Price from artificialanalysis.ai, 18 Oct 2026.
Benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| Artificial Analysis Speech Arena: Text to Speech (provider voices) | 1,042 | Higgs Audio V3 TTSProvider Voice Arena, all categories and accents; rank 68 of 96, 2159 appearances, 95% CI ±13; raw value from the page's embedded data (shown rounded); no last-updated date shown, so as_of = retrieval date. | 18 Oct 2026 | artificialanalysis.ai |
Where you can use it
No tracked tool lists this model yet; as open weights it runs in local runtimes.