AI Paper Atlas

Research map / Benchmarks / DynaMath

DynaMath leaderboard

13 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on DynaMath. Best accuracy: 83.4 (MMFineReason-8B (SFT+RL), NeurIPS 2026).

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.

DynaMath · accuracy

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

#accuracy ↑MethodPaperTopic
183.4MMFineReason-8B (SFT+RL)
Qwen3-VL-8B-Instruct
MMFineReason: Closing the Multimodal Reasoning Gap via Open Data-Centric Methods
NeurIPS 2026 · Table 3
Multimodal LLMs
261.6PyVision-Image
Qwen2.5-VL-7B
PyVision-RL: Forging Open Agentic Vision Models via RL
ICML 2026 · Table 1
Multimodal LLMs
358.5Vision-G1
Qwen2.5-VL-7B
Vision-G1: Towards General Reasoning Vision-Language Models via Reinforcement Learning
AAAI 2026 · Table 2
Multimodal LLMs
458.3FAST-7B
Qwen2.5-VL-7B
Fast-Slow Thinking GRPO for Large Vision-Language Model Reasoning
NeurIPS 2025 · Main results table
Multimodal LLMs
557.2DeepEyesV2
Qwen2.5-VL-7B
DeepEyesV2: Toward Agentic Multimodal Model
ICLR 2026 · Table 3
Multimodal LLMs
656.3Vision-R1-7B
Qwen2.5-VL-7B
Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models
ICLR 2026 · Table 1
Multimodal LLMs
751.4IP-Merging
InternVL3-8B-Instruct
Can MLLMs Absorb Math Reasoning Abilities from LLMs as Free Lunch?
NeurIPS 2025 · Table 1
Instruction and fine-tuning
848Bagel-Canvas + AdaViG
Bagel-7B
Model Guides You How to Draw: Adaptive Visual Gating for Unified Multimodal Reasoning
NeurIPS 2026 · Table 1
Multimodal LLMs
946.2CogFlow-7B
Qwen2.5-VL-7B
CogFlow: Bridging Perception and Reasoning through Knowledge Internalization for Visual Mathematical Problem Solving
ICLR 2026 · text
Multimodal LLMs
1039.5R-4B-RL (A-T)
R-4B
R-4B: Incentivizing General-Purpose Auto-Thinking in MLLMs via Bi-Mode Annealing and Reinforce Learning
CVPR 2026 · Table 1
Reasoning and chain of thought
3 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.