Open LMM Reasoning - DynaMath - Subject-puzzle test; Avg: leaderboard
The "DynaMath - Subject-puzzle test; Avg" subtask of Open LMM Reasoning. A task from the Open LMM Reasoning leaderboard, which evaluates large multimodal models on multi-step visual, mathematical and logical reasoning problems.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 100 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 Mini | 72.9 | #355 |
| 2 | GPT-4.1 (2025-04-14) | 67.6 | #250 |
| 3 | Doubao-1.5-Pro | 65.3 | #287 |
| 4 | Claude 3.7 Sonnet | 61.8 | #249 |
| 5 | GPT-4o ChatGPT | 61.8 | |
| 6 | Gemini 2.0 Flash | 57.6 | #342 |
| 7 | Claude 3.5 Sonnet (20241022) | 54.7 | #291 |
| 8 | QVQ-72B-Preview | 54.1 | #440 |
| 9 | InternVL3-38B | 52.9 | #405 |
| 10 | InternVL3-78B | 50.6 | #352 |
| 11 | InternVL3-14B | 49.4 | #505 |
| 12 | Gemini 1.5 Pro (002) | 48.8 | #349 |
| 13 | InternVL2.5-78B-CoT | 48.8 | #469 |
| 14 | GPT-4.1 Nano | 43.5 | #731 |
| 15 | Grok 2 (1212) | 42.9 | #496 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-lmm-reasoning-dynamath-subject-puzzle-test-avg · How It Works · Data refreshed daily, snapshot 2026-10-20.