Nari Qwen3-TTS and Qwen3-ASR

About Nari Qwen3-TTS and Qwen3-ASR

Nari ranks first in STT latency, second in STT WER, second in TTS latency, and first in TTS WER in Coval’s September 14, 2026 benchmarks.

In the maker’s words

Hey HN, Toby from Nari Labs here. We've been working on making OSS speech models super-fast. Last year, we built Dia, the first OSS text-to-speech model capable of doing natural dialogue. Since then, so many more great speech models have been released to the public. But the market is still dominated by closed source models. We think that's an inference problem. Existing systems such as vLLM / SGLang are not well suited for multimodal inference. To prove this, we built an inference engine specialized for Qwen3-TTS and open-sourced it ( https://github.com/nari-labs/nari-qwen3-tts ). Running at sub-50 ms latency at 10 RPS, this showed open models can be run much faster and cheaper. Since then, we've been working hard to bring cheap, fast, and high quality serving to all. And we've even beat closed models at their game! Measured on the highly cited Coval (YC S24) voice AI benchmarks, our Qwe…

Where people found it

• Hacker NewsShow HN: Nari Qwen3-TTS and Qwen3-ASR – High accuracy, low latency and cost92 points31 comments19 days ago
• Hacker NewsExpanding the Pareto Frontier for Realtime Transcription1 points1 comments22 days ago
• Hacker NewsNari Labs: Multimodal inference at the speed of light2 points0 comments23 days ago
• Hacker NewsShow HN: 10x cheaper TTS at 50ms time-to-first-audio1 points1 comments23 days ago