AI for Education Visual Maths - Number and Operations: leaderboard
This is the number and operations slice of Fab AI's Visual Maths Benchmark, a set of 237 early grade maths questions that each depend on an image, selected from the Math-Vision (Math-V) dataset. It scores only that strand, as the percent of questions answered correctly, where higher is better. Source: benchmarks.ai-for-education.org/leaderboards/maths.
Metric: Accuracy (%). Source: www.fab-ai.org. Saturation forecast: Around January 2027. 76 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 89.19 | #19 |
| 2 | Claude Fable 5 | 86.49 | #16 |
| 3 | GPT-5.5 | 83.78 | #31 |
| 4 | Gemini 3.1 Pro (Preview) | 81.08 | #58 |
| 5 | Gemini 3.7 Flash | 81.08 | #23 |
| 6 | Claude Opus 5 | 75.68 | #12 |
| 7 | Qwen 3.5 Plus | 72.97 | #125 |
| 8 | GPT-5.4 Mini (Medium) | 72.97 | #212 (GPT-5.4 Mini) |
| 9 | Gemini 3 Pro | 70.27 | #80 |
| 10 | Gemini 3.5 Flash | 70.27 | #60 |
| 11 | Gemini 3.6 Flash | 70.27 | #51 |
| 12 | GPT-5.4 (Medium) | 70.27 | #79 (GPT-5.4) |
| 13 | Qwen 3.6 35B A3B | 67.57 | #226 |
| 14 | GPT-5.2 (Medium) | 67.57 | #110 (GPT-5.2) |
| 15 | Qwen 3.5 Flash | 67.57 | #232 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ai-for-education-visual-maths-number-and-operations · How It Works · Data refreshed daily, snapshot 2026-10-20.