Tencent and Tsinghua Open-Source WeMM-Video Omni-Modal Video Models

Tencent WeChat Vision and Tsinghua released WeMM-Video, an open-source family of bilingual omni-modal foundation models in 2B to 35B-A3B sizes that jointly understands text, image, audio, and video, with the paper available on arXiv.

2026-11-20 ~ 2026-11-20 · 2 related posts