Open-weight & owned stack · Open lab releases

Tencent Releases AuK-Flash Speech Model Weights

Sep 12, 2026

Tencent published official weights for the distilled 1.5-billion-parameter AuK-Flash speech model on Hugging Face.

Read the original at reddit.comOpens the publisher's site in a new tab

More in Open-weight & owned stack

Z.ai publishes GLM-5.3 open weightsSep 8 MiniMax H3 highlighted as open-weight multimodal modelSep 7 InternLM releases 397B multimodal Intern-S2 modelSep 13 DeepSeek Model Nearly Matches GPT-6 Astra DesignSep 10 DeepSeek presents V4 FlashSep 8