# VideoMMMU → MathVision

Calibration · https://benchgap.net/calibration/1642 · data retrieved 2026-10-29

MathVision is estimated from VideoMMMU with a Michaelis–Menten curve fitted on 10 models measured on both: y = 2.0000·x / (1.14920 + x), R² = 0.77, cross-validated error 6.2 pp. It is used for 0 estimates.

## Fitted on

| Model | VideoMMMU | MathVision |
|---|---|---|
| [Claude Opus 4.5](https://benchgap.net/model/claude-opus-4-5.md) | 84.4% | 74.3% |
| [dots3-note Preview](https://benchgap.net/model/dots3-note-preview.md) | 86.8% | 87.7% |
| [Gemini 3 Pro](https://benchgap.net/model/gemini-3-pro.md) | 87.6% | 86.6% |
| [GPT-5.2](https://benchgap.net/model/gpt-5-2.md) | 85.9% | 83.0% |
| [Kimi K2.5](https://benchgap.net/model/kimi-k2-5.md) | 86.6% | 84.2% |
| [LFM3-VL-2B](https://benchgap.net/model/lfm3-vl-2b.md) | 49.6% | 51.6% |
| [Qwen3.5 397B](https://benchgap.net/model/qwen3-5-397b.md) | 84.7% | 88.6% |
| [Qwen3.6 Plus](https://benchgap.net/model/qwen3-6-plus.md) | 84.0% | 88.0% |
| [Qwen3.7 Plus](https://benchgap.net/model/qwen3-7-plus.md) | 85.4% | 90.3% |
| [Qwen3.8 Max](https://benchgap.net/model/qwen3-8-max.md) | 88.7% | 95.2% |