# VideoMMMU → CharXiv

Calibration · https://benchgap.net/calibration/1639 · data retrieved 2026-10-29

CharXiv is estimated from VideoMMMU with a Hill curve fitted on 10 models measured on both: y = 0.0000 + (1.2000 − 0.0000)·x^6.00 / (0.75823^6.00 + x^6.00), R² = 0.35, cross-validated error 5.9 pp. It is used for 0 estimates.

## Fitted on

| Model | VideoMMMU | CharXiv |
|---|---|---|
| [Claude Opus 4.5](https://benchgap.net/model/claude-opus-4-5.md) | 84.4% | 68.5% |
| [Gemini 3 Pro](https://benchgap.net/model/gemini-3-pro.md) | 87.6% | 81.4% |
| [GPT-5.2](https://benchgap.net/model/gpt-5-2.md) | 85.9% | 82.1% |
| [Kimi K2.5](https://benchgap.net/model/kimi-k2-5.md) | 86.6% | 77.5% |
| [Qwen3.5 397B](https://benchgap.net/model/qwen3-5-397b.md) | 84.7% | 80.8% |
| [Qwen3.6-27B](https://benchgap.net/model/qwen3-6-27b.md) | 84.4% | 78.4% |
| [Qwen3.6-35B-A3B](https://benchgap.net/model/qwen3-6-35b-a3b.md) | 83.7% | 78.0% |
| [Qwen3.6 Plus](https://benchgap.net/model/qwen3-6-plus.md) | 84.0% | 81.5% |
| [Qwen3.7 Plus](https://benchgap.net/model/qwen3-7-plus.md) | 85.4% | 85.9% |
| [Qwen3.8 Max](https://benchgap.net/model/qwen3-8-max.md) | 88.7% | 93.5% |