# MMLU-Redux → SuperGPQA

Calibration · https://benchgap.net/calibration/mmluredux/current/to/supergpqa/current · data retrieved 2026-11-18

SuperGPQA is estimated from MMLU-Redux with a Hill curve fitted on 9 models measured on both: y = 0.0000 + (1.2000 − 0.0000)·x^6.00 / (0.89960^6.00 + x^6.00), R² = 0.96, cross-validated error 4.1 pp. It is used for 6 estimates.

| Estimated model | MMLU-Redux | SuperGPQA | Source |
|---|---|---|---|
| [DeepSeek-R1-0528](https://benchgap.net/model/deepseek-r1-0528.md) | 93.4% | 66.7% | estimated ± 4.1 pp, high confidence |
| [Ternary Bonsai 2 27B](https://benchgap.net/model/ternary-bonsai-2-27b.md) | 89.1% | 58.3% | estimated ± 4.1 pp, high confidence |
| [Mellum2.1-12B-A2.5B-Thinking](https://benchgap.net/model/mellum2-1-12b-a2-5b-thinking.md) | 87.8% | 55.6% | estimated ± 4.1 pp, high confidence |
| [Mellum2-12B-A2.5B-Thinking](https://benchgap.net/model/mellum2-12b-a2-5b-thinking.md) | 86.2% | 52.4% | estimated ± 4.1 pp, high confidence |
| [Qwen3 32B](https://benchgap.net/model/qwen3-32b.md) | 85.7% | 51.3% | estimated ± 4.1 pp, high confidence |
| [Mellum2-12B-A2.5B-Instruct](https://benchgap.net/model/mellum2-12b-a2-5b-instruct.md) | 78.1% | 36.0% | estimated ± 4.1 pp, high confidence |

## Fitted on

| Model | MMLU-Redux | SuperGPQA |
|---|---|---|
| [Claude Opus 4.5](https://benchgap.net/model/claude-opus-4-5.md) | 96.6% | 70.6% |
| [Kimi K2.5](https://benchgap.net/model/kimi-k2-5.md) | 94.5% | 69.2% |
| [MiniCPM5-1B](https://benchgap.net/model/minicpm5-1b.md) | 70.1% | 23.1% |
| [MiniCPM5-2B](https://benchgap.net/model/minicpm5-2b.md) | 84.7% | 40.8% |
| [Qwen3.5 397B](https://benchgap.net/model/qwen3-5-397b.md) | 94.9% | 70.4% |
| [Qwen3.6-27B](https://benchgap.net/model/qwen3-6-27b.md) | 93.5% | 66.0% |
| [Qwen3.6 Plus](https://benchgap.net/model/qwen3-6-plus.md) | 94.5% | 71.6% |
| [Qwen3.7 Max](https://benchgap.net/model/qwen3-7-max.md) | 95.0% | 73.6% |
| [Qwen3.7 Plus](https://benchgap.net/model/qwen3-7-plus.md) | 94.5% | 71.4% |
