Tencent Hunyuan open-sources AuK audio model, drawing speed raves

bingxu_ · x · 2026-09-23

Tencent Hunyuan has open-sourced AuK, an audio generation model that users describe as remarkably fast for audio generation and editing — "almost like magic."

The post also raises an open question: how do open-source speech generation models now compare in quality with closed-source offerings like ElevenLabs? Rapid iteration on open models is narrowing the gap with commercial solutions.

Original post →

More from Multimodal

Multimodal channel →