Tencent Hunyuan Releases Next-Gen HyASR3.0 Speech Recognition Model

Tencent Hunyuan has launched its new HyASR3.0 preview speech recognition model. Built on the Hy3 LLM with a MoE architecture, it reduces the multilingual word error rate to approximately 3% while integrating high-precision recognition with deep semantic understanding.

2026-08-04 ~ 2026-08-04 · 2 related posts

• Tencent Hunyuan Launches HyASR3.0: Reduces Multilingual WER to Around 3% — 腾讯混元 · 2026-08-04
• Tencent Hunyuan Launches HyASR3.0 Speech Model, Reducing Multilingual WER to ~3% — 机器之心 · 2026-08-04