78.2%
Exact match
Run details
- Date
- 2026-06-05
- Test cases
- 11,759
- Median latency
- 7.77s
- Total cost
- $4.64
- Avg. prompt tokens
- 1,700.6
- Avg. answer tokens
- 662.1
- Avg. reasoning tokens
- —
- Quantization
- bf16
About this benchmark
German MMLU-ProX across 14 subjects; primary metric accuracy. Legacy rows have 11,759 items. Corrected v1.1 rows have 11,737 after a frozen, model-independent manifest excludes 20 duplicated-gold translation collisions, one independently reviewed ambiguous source item, and one independently confirmed notation-translation collapse; the row's item count identifies the definition.
Subject breakdown
| Subject | Score |
|---|---|
| biology | 87.3% |
| business | 83.9% |
| chemistry | 85.4% |
| computer science | 83.9% |
| economics | 82.6% |
| engineering | 73.5% |
| health | 71.2% |
| history | 66.7% |
| law | 50.7% |
| math | 92.0% |
| other | 67.5% |
| philosophy | 69.1% |
| physics | 86.1% |
| psychology | 78.6% |
Provenance
- Canonical model
- google/gemma-4-26b-a4b-it
- Run ID
2026-06-05T10-17-05__openrouter__google__gemma-4-26b-a4b-it__mmlu_prox_de