Vero : An Open RL Recipe for General Visual Reasoning

Gabriel Sarch ∗ Linrong Cai ∗ Qunzhong Wang Haoyang Wu Danqi Chen Zhuang Liu † Princeton University · 2026

Read Vero : An Open RL Recipe for General Visual Reasoning by Gabriel Sarch ∗ Linrong Cai ∗ Qunzhong Wang Haoyang Wu Danqi Chen Zhuang Liu † Princeton University (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-04917.txt.

Description

Sarch and five colleagues ask what it takes to build a visual reasoner for charts, science, spatial tasks, and open-ended questions. Leading vision-language models show visual reasoning is within reach, but their closed data and pipelines block study and reuse. Their Vero family of open models matches or beats open-weight models across visual reasoning tasks. They scale data and rewards across six task groups, building Vero-600K from 59 datasets with task-routed rewards. On their 30-benchmark VeroEval suite, Vero-600K beats datasets under controlled comparisons.

Catalog record

CREATORGabriel Sarch ∗ Linrong Cai ∗ Qunzhong Wang Haoyang Wu Danqi Chen Zhuang Liu † Princeton University
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE127467
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.04917

Browse the full library · About the Tower of Babel