# MathVision → PerceptionBench

Calibration · https://benchgap.net/calibration/mathvision/current/to/perceptionbench/current · data retrieved 2026-11-21

PerceptionBench is estimated from MathVision with a inverse Michaelis–Menten curve fitted on 5 models measured on both: y = 0.70017·(x − 0.0000) / (0.0000 + 2.0000 − x), R² = 0.58, cross-validated error 3.6 pp. It is used for 12 estimates.

| Estimated model | MathVision | PerceptionBench | Source |
|---|---|---|---|
| [Qwen3.8-Omni-Flash](https://benchgap.net/model/qwen3-8-omni-flash.md) | 91.8% | 59.4% | estimated ± 3.6 pp, medium confidence |
| [Qwen3.8-Flash-Next](https://benchgap.net/model/qwen3-8-flash-next.md) | 90.6% | 58.0% | estimated ± 3.6 pp, medium confidence |
| [Qwen3.7 Plus](https://benchgap.net/model/qwen3-7-plus.md) | 90.3% | 57.6% | estimated ± 3.6 pp, medium confidence |
| [Qwen3.5 397B](https://benchgap.net/model/qwen3-5-397b.md) | 88.6% | 55.7% | estimated ± 3.6 pp, medium confidence |
| [Qwen3.6 Plus](https://benchgap.net/model/qwen3-6-plus.md) | 88.0% | 55.0% | estimated ± 3.6 pp, medium confidence |
| [Kimi K2.6](https://benchgap.net/model/kimi-2-6.md) | 87.4% | 54.3% | estimated ± 3.6 pp, low confidence |
| [Gemini 3 Pro](https://benchgap.net/model/gemini-3-pro.md) | 86.6% | 53.5% | estimated ± 3.6 pp, low confidence |
| [Qwen3.5-122B-A10B](https://benchgap.net/model/qwen3-5-122b-a10b.md) | 86.2% | 53.0% | estimated ± 3.6 pp, low confidence |
| [Qwen3.5-27B](https://benchgap.net/model/qwen3-5-27b.md) | 86.0% | 52.8% | estimated ± 3.6 pp, low confidence |
| [Kimi K2.5](https://benchgap.net/model/kimi-k2-5.md) | 84.2% | 50.9% | estimated ± 3.6 pp, low confidence |
| [Qwen3.5-35B-A3B](https://benchgap.net/model/qwen3-5-35b-a3b.md) | 83.9% | 50.6% | estimated ± 3.6 pp, low confidence |
| [GPT-5.2](https://benchgap.net/model/gpt-5-2.md) | 83.0% | 49.7% | estimated ± 3.6 pp, low confidence |

## Fitted on

| Model | MathVision | PerceptionBench |
|---|---|---|
| [dots3-note Preview](https://benchgap.net/model/dots3-note-preview.md) | 87.7% | 53.4% |
| [Inkling 2 Mini](https://benchgap.net/model/inkling-2-mini.md) | 89.0% | 57.5% |
| [Inkling 2 Small](https://benchgap.net/model/inkling-2-small.md) | 92.8% | 64.7% |
| [Kimi K3](https://benchgap.net/model/kimi-k3.md) | 94.3% | 58.5% |
| [Qwen3.8 Max](https://benchgap.net/model/qwen3-8-max.md) | 95.2% | 63.5% |
