Open LMM Reasoning - WeMath - Two-step(S2): leaderboard
The "WeMath - Two-step(S2)" 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. 83 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Doubao-1.5-Pro | 78.6 | #287 |
| 2 | GPT-4o ChatGPT | 69.2 | |
| 3 | GPT-4.1 (2025-04-14) | 68.9 | #250 |
| 4 | Claude 3.7 Sonnet | 68.6 | #249 |
| 5 | InternVL3-38B | 67.2 | #405 |
| 6 | Gemini 1.5 Pro (002) | 65.8 | #349 |
| 7 | InternVL3-78B | 62.8 | #352 |
| 8 | InternVL3-14B | 62.8 | #505 |
| 9 | Taichu-VLR-7B | 62.8 | #526 |
| 10 | Gemini 2.0 Flash | 61.1 | #342 |
| 11 | Claude 3.5 Sonnet (20241022) | 60.6 | #291 |
| 12 | VLAA-Thinker-Qwen2.5VL-7B | 60.3 | #555 |
| 13 | QVQ-72B-Preview | 59.2 | #440 |
| 14 | InternVL2.5-78B | 58.1 | #431 |
| 15 | InternVL3-8B | 55.8 | #620 |
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-wemath-two-step-s2 · How It Works · Data refreshed daily, snapshot 2026-10-20.