# VideoMMMU → ERQA

Calibration · https://benchgap.net/calibration/1641 · data retrieved 2026-10-29

ERQA is estimated from VideoMMMU with a Michaelis–Menten + offset curve fitted on 8 models measured on both: y = 0.1321 + 2.0000·x / (2.44830 + x), R² = 0.36, cross-validated error 9.6 pp. It is used for 0 estimates.

## Fitted on

| Model | VideoMMMU | ERQA |
|---|---|---|
| [Claude Opus 4.5](https://benchgap.net/model/claude-opus-4-5.md) | 84.4% | 46.8% |
| [Gemini 3 Pro](https://benchgap.net/model/gemini-3-pro.md) | 87.6% | 70.5% |
| [GPT-5.2](https://benchgap.net/model/gpt-5-2.md) | 85.9% | 59.8% |
| [LFM3-VL-2B](https://benchgap.net/model/lfm3-vl-2b.md) | 49.6% | 48.2% |
| [Qwen3.5 397B](https://benchgap.net/model/qwen3-5-397b.md) | 84.7% | 67.5% |
| [Qwen3.6-27B](https://benchgap.net/model/qwen3-6-27b.md) | 84.4% | 62.5% |
| [Qwen3.7 Plus](https://benchgap.net/model/qwen3-7-plus.md) | 85.4% | 69.8% |
| [Qwen3.8 Max](https://benchgap.net/model/qwen3-8-max.md) | 88.7% | 77.8% |