Starstruk Logo MarkStarstruk

v0.6

tanchongmin/strictjsonv0.6Aug 18, 2025by liPatrick

AI Summary

v0.6 introduces new model variants (gemma3 and qwen3), enhanced language understanding (specifically Hindi), and improved audio robustness. It also adds support for the VoiceBench evaluation suite.

Key Highlights

  • New gemma3 and qwen3 variants alongside base llama models.
  • Improved background noise robustness and audio quality across all languages.
  • New `<noise>` token handling for non-human or noisy audio.
  • New eval support for VoiceBench.

New Features

  • New gemma3 and qwen3 model variants.
  • Introduction of the `<noise>` token for non-human audio.
  • VoiceBench evaluation suite integration.
  • Allow response generation with no user message.

Full Release Notes

We're releasing **Ultravox v0.6** today. The [weights](https://huggingface.co/fixie-ai) have been pushed to Hugging Face. If you're using the [Ultravox Realtime APIs](https://docs.ultravox.ai), v0.6 is the new default.

## What's New
* v0.6 improves upon 0.5 in the following ways:
* Improvements on Hindi language understanding.
* Produces <noise> token on noisy or non-human audio. 
* Improved background noise and audio quality robustness in all languages. 
* New gemma3 and qwen3 variants in addition to our base llama 3.3 models.

## Evals
New eval support for [VoiceBench](https://github.com/MatthewCYM/VoiceBench). The benchmark tests speech-language models on 9 different tasks ranging from open-form text generation to question-answering and instruction following. 

## Training
This version of Ultravox continues to use a frozen Llama pre-trained core (3.1 for 8B and 3.3 for 70B), along with new gemma3 27b and qwen3 32b variants.

## What's Changed
Training Stability: Patch HF Hub and Datasets methods and update [datasets.py](http://datasets.py/) by @farzadab #280
General improvement. by @zqhuang211 #281
Ultravox v0.6 + general improvements by @liPatrick #309
Allow response generation with no user message by @matthewclso #310
Add voicebench evaluation suite by @zqhuang211 #312

**Full Changelog**: https://github.com/fixie-ai/ultravox/compare/v0.5...v0.6