AI for Education Visual Reasoning — leaderboard
Metric: Accuracy (%). Source: www.fab-ai.org. 64 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.5 Flash | 86 |
| 2 | GPT-5.6 Sol | 84.9 |
| 3 | Gemini 3 Flash | 78 |
| 4 | GPT-5.5 | 77.7 |
| 5 | GPT-5.4 (Medium) | 76.3 |
| 6 | Qwen 3.5 397B A17B | 76 |
| 7 | Gemini 3 Pro | 75.9 |
| 8 | Qwen 3.6 Plus | 73.3 |
| 9 | Qwen 3.5 Plus | 72.6 |
| 10 | GPT-5.2 (Medium) | 67.3 |
| 11 | Qwen 3.6 27B | 66.2 |
| 12 | GPT-5.4 Mini (Medium) | 64.2 |
| 13 | Qwen 3.6 35B A3B | 64.2 |
| 14 | qwen3.5-flash | 61.6 |
| 15 | Kimi K2.5 | 60.5 |
Interactive version: theaggregate.ai/benchmark?slug=ai-for-education-visual-reasoning · How the rankings work · Data refreshed daily, snapshot 2026-07-22.