v0.6
bradtraversy/design-resources-for-developersv0.6Aug 18, 2025by liPatrick
AI Summary
Updates Ultravox with improved language understanding (Hindi), enhanced noise robustness, and introduces new gemma3 and qwen3 model variants as the default for Realtime APIs.
Key Highlights
- Weights pushed to Hugging Face and set as default for Realtime APIs
- Introduces new gemma3 and qwen3 model variants
- Improves Hindi language understanding and background noise robustness
- Adds VoiceBench evaluation suite support
New Features
- New model variants (gemma3, qwen3)
- VoiceBench evaluation support
- Noise token generation for non-human audio
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