Individual experiment result
MMLU-ProX (DE)
Gemini 3.8 Flash Google
Exact match reasoning on / locked provider-internal
89.0%
Exact match
Run details
- Date
- 2026-09-04
- Test cases
- 11,759
- Median latency
- 7.24s
- Total cost
- $104.01
- Avg. prompt tokens
- 1,664.9
- Avg. answer tokens
- 222.4
- Avg. reasoning tokens
- 623.9
- Quantization
- provider-internal
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 | 93.7% |
| business | 92.3% |
| chemistry | 91.8% |
| computer science | 91.5% |
| economics | 91.1% |
| engineering | 86.8% |
| health | 82.4% |
| history | 80.8% |
| law | 76.9% |
| math | 95.4% |
| other | 84.6% |
| philosophy | 87.2% |
| physics | 93.5% |
| psychology | 89.2% |
Provenance
- Canonical model
- google/gemini-3.8-flash
- Run ID
2026-09-04T19-33-44__vertex_ai__gemini-3.8-flash__mmlu_prox_de