AI Paper Atlas

Research map / Benchmarks / MathVista

MathVista leaderboard

134 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on MathVista. Best accuracy: 85.2 (o4-mini + DreamPRM, NeurIPS 2025).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

MathVista · accuracy

101 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
185.2o4-mini + DreamPRM
o4-mini
DreamPRM: Domain-reweighted Process Reward Model for Multimodal Reasoning
NeurIPS 2025 · o4-mini PRM variants table
Reasoning and chain of thought
285Ours (+sampling)
Qwen3-VL-4B-Thinking
When to Think and When to Look: Uncertainty-Guided Lookback
CVPR 2026 · Table 2
Multimodal LLMs
383.7Agent0-VL-8B
Qwen3-VL-8B
Agent0-VL: Exploring Self-Evolving Agent for Tool-Integrated Vision-Language Reasoning
ICML 2026 · Table 1
Multimodal LLMs
482.9SV w/ GPT-4o Verdict
GPT-4o verdict
Small Drafts, Big Verdict: Information-Intensive Visual Reasoning via Speculation
ICLR 2026 · Table 2
Multimodal LLMs
582.1Octopus-8B
Qwen3-VL-8B-Instruct
Learning Self-Correction in Vision–Language Models via Rollout Augmentation
ICML 2026 · Table 2
Multimodal LLMs
681.2AnE-3rd
Qwen2.5-VL-7B-Instruct
AnE: Pushing the Reasoning Frontier of Multimodal LLMs via Anchor Evolution
ECCV 2026 · Table 1
Reasoning and chain of thought
780.94TTRV
MM-Eureka-7B
TTRV: Test-Time Reinforcement Learning for Vision Language Models
CVPR 2026 · Appendix table
Multimodal LLMs
880.7HICRA
MiMO-VL-Instruct-2508
Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning
ICLR 2026 · Table 2
Reasoning and chain of thought
979.9MiniCPM-V 4.5MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe
CVPR 2026 · Table 1
Multimodal LLMs
1078.5SRPO-32B
Qwen2.5-VL-32B-Instruct
SRPO: Enhancing Multimodal LLM Reasoning via Reflection-Aware Reinforcement Learning
NeurIPS 2025 · Table 1
Reasoning and chain of thought
91 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

MathVista · score

33 papers, each with its best reported number for its own method; higher is better.

#score ↑MethodPaperTopic
185.9UniDFlow (4B)Best of Both Worlds: Multimodal Reasoning and Generation via Unified Discrete Flow Matching
ICML 2026 · Table 1
Multimodal LLMs
284.3Wan-WeaverWan-Weaver: Interleaved Multi-modal Generation via Decoupled Training
CVPR 2026 · Table 2
Multimodal LLMs
374.7VLsI-7B
Qwen2-VL-7B
VLsI: Verbalized Layers-to-Interactions from Large to Small Vision Language Models
CVPR 2025 · Table 1
Multimodal LLMs
474.5Qwen2.5-VL-7B w. RIL
Qwen2.5-VL-7B
Unified Reinforcement and Imitation Learning for Vision-Language Models
NeurIPS 2025 · Table 1
Multimodal LLMs
574.3R1-VL-7B*
Qwen2.5-VL-7B
R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimization
ICCV 2025 · Table 1
Reasoning and chain of thought
673.3Uni-CoTUni-CoT: Towards Unified Chain-of-Thought Reasoning Across Text and Vision
ICLR 2026 · Table 3
Multimodal LLMs
773.2Qwen2-VL-7B+Critic-V
Qwen2-VL-7B
Critic-V: VLM Critics Help Catch VLM Errors in Multimodal Reasoning
CVPR 2025 · Main results table
Multimodal LLMs
873Qwen-PeRL
Qwen2.5-VL-7B-Instruct
PeRL: Permutation-Enhanced Reinforcement Learning for Interleaved Vision-Language Reasoning
NeurIPS 2025 · Main Results table
Multimodal LLMs
972.1OVR-7B
Qwen2.5-VL-7B
Open Vision Reasoner: Transferring Linguistic Cognitive Behavior for Visual Reasoning
NeurIPS 2025 · Visual reasoning benchmark tab
Multimodal LLMs
1070.4VLM-R³
Qwen2.5-VL 7B
VLM-R³: Region Recognition, Reasoning, and Refinement for Enhanced Multimodal Chain-of-Thought
NeurIPS 2025 · Main Result table
Multimodal LLMs
23 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.