Skip to content

DynaMath

A multimodal benchmark for dynamic mathematical reasoning over visual and structured inputs.

6 models

Higher is better.

  1. 1GPT-5.2OpenAI86.8
  2. 2Qwen3.5 397BAlibaba86.3
  3. 3Qwen3.6-27BAlibaba85.6
  4. 4Gemini 3 ProGoogle85.1
  5. 5Kimi K2.5Moonshot AI84.4
  6. 6Claude Opus 4.5Anthropic79.7
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Multimodal mathematical reasoning
Tasks
Dynamic visual math problems
Authors
Qwen (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.