Tutorials & Tips

Boson AI and LMSYS release end-to-end Higgs Audio v3 TTS service on SGLang-Omni

Original title:Blog Higgs Audio v3 TTS on SGLang-Omni: Real-Time, Controllable Speech for Voice Agents Today we are announcing end-to-end serving for Higgs Audio v3 TTS on SGLang-Omni. Higgs Audio v3 TTS is Boson AI's text-to-speech model for conversational voice agents: it generates natural and expres... Boson AI & SGLang-Omni Team

AI summary

Boson AI and LMSYS jointly launch an end-to-end Higgs Audio v3 TTS service built on the SGLang-Omni inference framework. The model has about 4B parameters, is based on the Qwen3-4B backbone, supports 100 languages (internal evaluation covers 111), and achieves character-level WER/CER on Seed-TTS, CV3, MiniMax-Multilingual and Higgs-Multilingual zero-shot voice cloning tasks. Developers can adjust emotion (20+ types), style, prosody (speed/pitch/pauses) and sound effects in real time via in-text control tags. The model supports streaming synthesis, starting speech generation before the text is complete while maintaining consistency. SGLang-Omni is designed for multi-stage generative models, unifying AR decoding and lightweight computation scheduling for low-latency inference.

Key Takeaways

  • Boson AI and LMSYS jointly launch an end-to-end Higgs Audio v3 TTS service built on the…
  • The model has about 4B parameters,is based on the Qwen3-4B backbone,supports 100 languages…
  • Developers can adjust emotion (20+ types),style,prosody (speed/pitch/pauses) and sound effects in…
  • The model supports streaming synthesis,starting speech generation before the text is complete…
  • SGLang-Omni is designed for multi-stage generative models,unifying AR decoding and lightweight…

Quick Facts

Published
Updated
Source
LMSYS Blog
Category
Tutorials & Tips
AI score
77
Read the original Read the full article at the original source
Why it matters

Higgs v3 把多语言和实时可控制整合进一个流式 TTS 模型,SGLang-Omni 为它搭好了多阶段推理的底座,做语音 Agent 的可以直接抄作业,这比多数论文落地快半拍。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: Boson AI and LMSYS release end-to-end Higgs Audio v3 TTS service on SGLang-Omni. June 4, 2026. https://new.nat6.net/en/news/cmpzqhv5y0607slkptes62n17/

View all →