Open LMM Reasoning - WeMath - Understanding of Solid Figures: leaderboard

Metric: Accuracy (%). Source: huggingface.co. 83 models tracked.

Top models

#ModelScore
1Gemini 1.5 Pro (002)74.6
2Doubao-1.5-Pro73.1
3Claude 3.7 Sonnet71.4
4GPT-4.1 (2025-04-14)70.5
5InternVL3-38B68.2
6GPT-4o ChatGPT68.1
7VLAA-Thinker-Qwen2.5VL-7B67.7
8Gemini 2.0 Flash67.4
9InternVL3-8B67.1
10InternVL2.5-8B-BoN-866.2
11Taichu-VLR-7B66.1
12QVQ-72B-Preview65.7
13Claude 3.5 Sonnet (20241022)64.3
14URSA-8B62.9
15Gemma 3 27B61.9

Interactive version: theaggregate.ai/benchmark?slug=open-lmm-reasoning-wemath-understanding-of-solid-figures · How It Works · Data refreshed daily, snapshot 2026-09-05.