Open LMM Reasoning - WeMath - Direction — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 83 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4.1 (2025-04-14) | 96.4 |
| 2 | GPT-4o ChatGPT | 96.4 |
| 3 | Doubao-1.5-Pro | 96.4 |
| 4 | Grok 2 (1212) | 92.9 |
| 5 | Claude 3.7 Sonnet | 89.8 |
| 6 | InternVL3-78B | 89.5 |
| 7 | InternVL3-8B | 86.4 |
| 8 | InternVL3-14B | 86.4 |
| 9 | Gemini 1.5 Pro (002) | 86.2 |
| 10 | InternVL2.5-78B | 86.2 |
| 11 | QVQ-72B-Preview | 85.7 |
| 12 | Gemini 2.0 Flash | 82.9 |
| 13 | InternVL3-38B | 82.6 |
| 14 | Claude 3.5 Sonnet (20241022) | 82.6 |
| 15 | Gemma 3 12B | 82.6 |
Interactive version: theaggregate.ai/benchmark?slug=open-lmm-reasoning-wemath-direction · How the rankings work · Data refreshed daily, snapshot 2026-07-22.