Open LMM Reasoning - DynaMath - Subject-arithmetic: leaderboard
The "DynaMath - Subject-arithmetic" 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 | 57.7 | #355 |
| 2 | GPT-4o ChatGPT | 53.8 | |
| 3 | Doubao-1.5-Pro | 50 | #287 |
| 4 | GPT-4.1 (2025-04-14) | 46.2 | #250 |
| 5 | Gemini 2.0 Flash | 38.5 | #342 |
| 6 | Taichu-VLR-7B | 38.5 | #526 |
| 7 | Claude 3.5 Sonnet (20241022) | 34.6 | #291 |
| 8 | InternVL3-8B | 34.6 | #620 |
| 9 | Claude 3.7 Sonnet | 34.6 | #249 |
| 10 | GPT-4.1 Nano | 30.8 | #731 |
| 11 | Gemini 1.5 Pro (002) | 30.8 | #349 |
| 12 | InternVL3-78B | 30.8 | #352 |
| 13 | InternVL3-14B | 30.8 | #505 |
| 14 | Gemma 3 27B | 30.8 | #607 |
| 15 | Moonshot v1 8K | 30.8 | #617 |
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-arithmetic · How It Works · Data refreshed daily, snapshot 2026-10-20.