MyPrivateClaw

Tencent's Hunyuan lab open-sources AuK, a foundational model for speech generation and editing

Tencent's Hunyuan lab open sourced AuK, an open source foundational model for unified speech generation and editing, on September 9, 2026, with source code and…

Published on MyPrivateClaw

Sep 12, 2026, 7:41 PM UTC

Coverage date

Sep 12, 2026

Last updated

Sep 12, 2026, 7:41 PM UTC

News summary

Tencent's Hunyuan lab open sourced AuK, an open source foundational model for unified speech generation and editing, on September 9, 2026. According to the release, source code and model weights are publicly available. A distilled AuK Flash variant performs 4 step inference without classifier free guidance and achieves a 4.5 wall clock speedup over the full AuK model under matched conditions, per the paper's stated benchmark attribution. AuK is released under the MIT License, and SGLang Omni provided Day 0 support for AuK and AuK Flash as of the September 9, 2026 open source announcement.