2026-07-02 Teaching Vision-Language-Action Models What to See and Where to Look Yuguang Yang et.al. 2607.01658 link 2026-07-02 VLAFlow: A Unified Training Framework for Vision-Language-Action Models ...

Discrete diffusion LMs can draft radiology reports interactively - and match autoregression while doing it. We finetune an MoE diffusion VLM (DiffusionGemma-26B, 3.8B active) head-to-head against its ...