AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

ProJudge: A Multi-Modal Multi-Discipline Benchmark and Instruction-Tuning Dataset for MLLM-based Process Judges

Jiaxin Ai, Pengfei Zhou, Zhaopan Xu, Ming Li, Fanrui Zhang, Zizhen Li, Jianwen Sun, Yukang Feng, Baojin Huang, Zhongyuan Wang, Kaipeng Zhang

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Unified any-modality language models · 7 citations (OpenAlex)

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
scale InternVL2.5 judge from 8B to 38BInternVL2.5 · ProJudgeBenchstep correctness accuracy: 25.58 -> 78.85Improved
fine-tune on ProJudge-173k with Dynamic Dual-PhaseQwen2.5-VL-3B · ProJudgeBenchstep correctness accuracy: 11.47 -> 81.29Improved
fine-tune on ProJudge-173k with Dynamic Dual-PhaseInternVL2.5-8B · ProJudgeBenchstep correctness accuracy: 25.58 -> 84.50Improved
add Dynamic Dual-Phase to standard fine-tuning (ablation)InternVL2.5-8B · ProJudgeBench (overall)step correctness accuracy: 83.37 -> 84.50Improved
add Dynamic Dual-Phase to standard fine-tuning (ablation)Qwen2.5-VL-3B · ProJudgeBench (overall)step correctness accuracy: 80.57 -> 81.29Improved
add Dynamic Dual-Phase to standard fine-tuning (ablation)Qwen2.5-VL-7B · ProJudgeBench (overall)step correctness accuracy: 81.91 -> 83.72Improved
add Dynamic Dual-Phase to standard fine-tuning (ablation)Qwen2.5-VL-7B · ProJudgeBench OlympicArena subset (unseen)step correctness accuracy: 77.88 -> 82.67Improved
add Dynamic Dual-Phase to standard fine-tuning (ablation)InternVL2.5-8B · ProJudgeBench OlympicArena subset (unseen)step correctness accuracy: 81.71 -> 85.07Improved
add Dynamic Dual-Phase to standard fine-tuning (ablation)Qwen2.5-VL-3B · ProJudgeBench OlympicArena subset (unseen)step correctness accuracy: 80.38 -> 81.01Improved
increase problem difficulty from high school to OlympiadBenchGemini-2.0 · ProJudgeBenchstep correctness accuracy: 80.2 -> 56.2Hurt
judge solutions from a weak student model versus a strong one (GPT-4o judging MiniCPM-V-2_6 vs reverse)GPT-4o / MiniCPM-V-2_6 as judges · ProJudgeBench cross-model evaluationjudge accuracy: 84.31 vs 15.69Mixed
evaluate biology versus physics problemsQwen2.5-VL-72B · ProJudgeBench by disciplinestep correctness accuracy: 84.13 (biology) vs 77.08 (physics)Mixed

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