# MMMLU → MMLU-Redux

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

MMLU-Redux is estimated from MMMLU with a offset logistic curve fitted on 7 models measured on both: y = 0.0002 + (0.9667 − 0.0002) / (1 + exp(−14.08·(x − 0.6191))), R² = 0.99, cross-validated error 0.5 pp. It is used for 7 estimates.

| Estimated model | MMMLU | MMLU-Redux | Source |
|---|---|---|---|
| [Interfaze Beta](https://benchgap.net/model/interfaze-beta.md) | 90.9% | 95.1% | estimated ± 0.5 pp, low confidence |
| [Claude 4 Opus Thinking](https://benchgap.net/model/claude-4-opus-thinking.md) | 88.8% | 94.5% | estimated ± 0.5 pp, medium confidence |
| [Claude 4 Opus](https://benchgap.net/model/claude-4-opus.md) | 87.4% | 94.1% | estimated ± 0.5 pp, medium confidence |
| [K-EXAONE 2.0](https://benchgap.net/model/k-exaone-2-0.md) | 86.6% | 93.8% | estimated ± 0.5 pp, medium confidence |
| [Claude 3.7 Sonnet Thinking](https://benchgap.net/model/claude-3-7-sonnet-thinking.md) | 86.1% | 93.6% | estimated ± 0.5 pp, medium confidence |
| [Gemma 4 12B](https://benchgap.net/model/gemma-4-12b.md) | 83.4% | 92.2% | estimated ± 0.5 pp, medium confidence |
| [Claude 3.7 Sonnet](https://benchgap.net/model/claude-3-7-sonnet.md) | 83.2% | 92.1% | estimated ± 0.5 pp, medium confidence |

## Fitted on

| Model | MMMLU | MMLU-Redux |
|---|---|---|
| [Qwen3 235B-A22B](https://benchgap.net/model/qwen3-235b-a22b.md) | 79.8% | 89.2% |
| [Qwen3 235B-A22B (Reasoning)](https://benchgap.net/model/qwen3-235b-a22b-reasoning.md) | 84.3% | 92.7% |
| [Qwen3 32B](https://benchgap.net/model/qwen3-32b.md) | 76.5% | 85.7% |
| [Qwen3 32B (Reasoning)](https://benchgap.net/model/qwen3-32b-reasoning.md) | 80.6% | 90.9% |
| [Qwen3.7 Max](https://benchgap.net/model/qwen3-7-max.md) | 90.3% | 95.0% |
| [Qwen3.7 Plus](https://benchgap.net/model/qwen3-7-plus.md) | 89.0% | 94.5% |
| [QwQ-32B](https://benchgap.net/model/qwq-32b.md) | 80.9% | 90.0% |
