Open LMM Reasoning - LogicVista - deductive: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 83 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Doubao-1.5-Pro | 88.2 |
| 2 | GPT-4o ChatGPT | 86 |
| 3 | InternVL3-38B | 84.9 |
| 4 | Claude 3.5 Sonnet (20241022) | 83.9 |
| 5 | InternVL3-78B | 81.7 |
| 6 | Gemini 1.5 Pro (002) | 81.7 |
| 7 | GPT-4.1 (2025-04-14) | 81.7 |
| 8 | QVQ-72B-Preview | 81.7 |
| 9 | GPT-4.1 Mini | 80.6 |
| 10 | InternVL2.5-78B | 75.3 |
| 11 | Claude 3.7 Sonnet | 74.2 |
| 12 | Gemma 3 27B | 74.2 |
| 13 | Grok 2 (1212) | 69.9 |
| 14 | Gemma 3 12B | 69.9 |
| 15 | Gemini 2.0 Flash | 68.8 |
Interactive version: theaggregate.ai/benchmark?slug=open-lmm-reasoning-logicvista-deductive · How It Works · Data refreshed daily, snapshot 2026-09-05.