# ScreenSpot Pro → MathVision

Calibration · https://benchgap.net/calibration/1414 · data retrieved 2026-10-29

MathVision is estimated from ScreenSpot Pro with a Michaelis–Menten curve fitted on 7 models measured on both: y = 2.0000·x / (0.95298 + x), R² = 0.51, cross-validated error 12.4 pp. It is used for 4 estimates.

| Estimated model | ScreenSpot Pro | MathVision | Source |
|---|---|---|---|
| [Holo2-235B-A22B](https://benchgap.net/model/holo2-235b-a22b.md) | 70.6% | 85.1% | estimated ± 12.4 pp, low confidence |
| [Holo2-30B-A3B](https://benchgap.net/model/holo2-30b-a3b.md) | 66.1% | 81.9% | estimated ± 12.4 pp, low confidence |
| [Holo2-8B](https://benchgap.net/model/holo2-8b.md) | 58.9% | 76.4% | estimated ± 12.4 pp, low confidence |
| [Holo2-4B](https://benchgap.net/model/holo2-4b.md) | 57.2% | 75.0% | estimated ± 12.4 pp, low confidence |

## Fitted on

| Model | ScreenSpot Pro | MathVision |
|---|---|---|
| [Claude Opus 4.5](https://benchgap.net/model/claude-opus-4-5.md) | 45.7% | 74.3% |
| [Gemini 3 Pro](https://benchgap.net/model/gemini-3-pro.md) | 72.7% | 86.6% |
| [LFM3-VL-2B](https://benchgap.net/model/lfm3-vl-2b.md) | 55.5% | 51.6% |
| [Qwen3.5 397B](https://benchgap.net/model/qwen3-5-397b.md) | 65.6% | 88.6% |
| [Qwen3.6 Plus](https://benchgap.net/model/qwen3-6-plus.md) | 68.2% | 88.0% |
| [Qwen3.7 Plus](https://benchgap.net/model/qwen3-7-plus.md) | 79.0% | 90.3% |
| [Qwen3.8 Max](https://benchgap.net/model/qwen3-8-max.md) | 84.5% | 95.2% |