WeChat Team Releases 31.7B Unified Image Model WeLM-Image, Weights Not Open

WeChat's team introduced WeLM-Image, a 31.7B-parameter (5.4B active) unified image foundation model using a MoT-MoE architecture, designed for single-GPU deployment. The model integrates multimodal understanding into text-to-image, instruction editing, and multi-reference synthesis, but its weights are not open-sourced.

2026-11-08 ~ 2026-11-08 · 2 related posts

• Tencent's WeLM-Image: unified 31.7B image model, weights not open-sourced — bdsqlsz · 2026-11-08
• WeChat's WeLM-Image paper: MoT-MoE architecture for single-GPU 31.7B image generation — bdsqlsz · 2026-11-08