AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Vero: Open Reinforcement Learning Recipes for Visual Reasoning

Gabriel Sarch, Linrong Cai, Qunzhong Wang, Haoyang Wu, Danqi Chen, Zhuang Liu

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Reinforcement learning for visual reasoning

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use RL instead of SFT on Vero-600K (ablation)Qwen2.5-VL-7B-Instruct · VeroEval overallavg score: 52.8 -> 57.2Improved
train SFT on Vero-600K instead of FineVision (ablation)Qwen2.5-VL-7B-Instruct · VeroEval overallavg score: 46.2 -> 52.8Improved
use task-routed reward instead of math_verify (ablation)Qwen2.5-VL-7B-Instruct · VeroEval overallavg score: 51.8 -> 57.2Improved
use task-routed reward instead of math_verify (ablation)Qwen2.5-VL-7B-Instruct · VeroEval Captioning & IFcategory avg: 34.3 -> 70.6Improved
use GSPO instead of GRPO (ablation)Qwen2.5-VL-7B-Instruct · VeroEval 5 categoriesavg score: 54.3 -> 54.7Improved
use GSPO instead of DAPO for entropy stability (ablation)Qwen2.5-VL-7B-Instruct · training entropyentropy: 0.22 -> 0.58Improved
use uniform task-category weighting instead of difficulty weighting (ablation)Qwen3-VL-8B-Instruct · VeroEval benchmark avg gainscore change: +5.2 -> +5.8Improved
include Knowledge & Recognition category in mixture (ablation)Qwen3-VL-8B-Instruct · VeroEval benchmark avg gainscore change: +4.9 -> +5.8Improved
add question filtering (ablation)Qwen2.5-VL-7B-Instruct · Spatial & Actioncategory avg: 56.3 -> 58.2Improved
add question filtering (ablation)Qwen2.5-VL-7B-Instruct · Grounding, Counting & Searchcategory avg: 54.7 -> 54.1Hurt
train on diverse task mixture instead of single category (Chart & OCR) (ablation)Qwen2.5-VL-7B-Instruct · Grounding, Counting & Searchscore change: -3.2 -> positiveImproved
select data with Vero filtering instead of random subset (ablation)Qwen2.5-VL-7B-Instruct · STEMavg benchmark gain: -0.2 -> +4.6Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy