AdaptMMBench - Auxiliary-Line Geometry: leaderboard
Metric: Accuracy (%) on the 120 geometry problems that need a self-generated auxiliary line, in adaptive mode (the model may draw on the image through tools); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 86.67 | #91 |
| 2 | Gemini 3 Pro | 78.33 | #77 |
| 3 | Qwen 3 VL 235B A22B Instruct | 68.33 | #264 |
| 4 | Qwen 3 VL 32B Instruct | 58.33 | #276 |
| 5 | Qwen 3 VL 8B Instruct | 46.67 | #401 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=adaptmmbench-auxiliary-line-geometry · How It Works · Data refreshed daily, snapshot 2026-10-11.