Liu Zhuang and Danqi Chen team open-source Vero, a general visual reasoning RL framework, reaching SOTA with zero thinking data

刘壮陈丹琦新作：开源通用视觉推理RL框架，0思考数据刷新SOTA

Princeton researchers including Liu Zhuang and Danqi Chen open-sourced Vero, an RL framework for visual reasoning, and report beating Qwen3-VL-8B-Thinking on 23 of 30 benchmarks. The post says Vero uses 600K samples filtered from 59 datasets, task-routed rewards, and single-stage RL across six task groups. The key point is the mechanism mix: no private thinking data, but the post does not disclose training cost or base model configuration.

Why it matters: Featured on HKR-H/K/R: the zero-thinking-data claim is a strong hook, and the post includes concrete benchmark and method details. I keep it in the low 80s because training cost, base model choice, and full reproduction conditions are not disclosed.