All models
flagship

Higgs Audio v2

Apache-2.0

Built on Llama 3.2 3B, 10M+ hours of training audio

Higgs Audio v2 builds audio generation on top of a Llama-3-class language model trained on 10M+ hours of audio. It wins the majority of emotion benchmarks against leading commercial models and does things few others can: multi-speaker dialogue, melodic humming, and voice cloning with deep prosody understanding.

Voices

1 voice
Neutral
Defaultdefault
—
NeutralEmotionalNumbers & Dates

The scripts

Neutral
“The quick brown fox jumps over the lazy dog near the riverbank.”
Emotional
“I can't believe you actually did it. This is incredible!”
Numbers & Dates
“On March 14th, 2025, the team raised $4.2 million at a 38% margin.”

Install

pip install higgs-audio

Higgs Audio v2 FAQ

Do I need a GPU to run Higgs Audio v2?

For practical speeds, yes — plan on roughly 12 GB of VRAM. If you'd rather not manage a GPU, hosted access to every model in this directory is coming via OpenSpeech Cloud.

Can Higgs Audio v2 clone voices?

Yes — Higgs Audio v2 supports voice cloning from reference audio.

What license is Higgs Audio v2 released under?

Higgs Audio v2 is released under the Apache-2.0 license. Always check the repository for the exact terms — some models license code and weights separately.

What languages does Higgs Audio v2 support?

Higgs Audio v2 supports English, Chinese, Spanish, French, German, Japanese, Korean.

Is there a hosted Higgs Audio v2 API?

Yes — OpenSpeech Cloud is launching hosted, pay-per-minute access to Higgs Audio v2 and every other model in this directory. No CUDA, Docker, or GPU needed.