Individual experiment result
MMLU-ProX (DE)
Gemini 3 Flash Preview Google
Exact match reasoning off provider-internal
86.3%
Exact match
Run details
- Date
- 2026-06-25
- Test cases
- 11,759
- Median latency
- 3.26s
- Total cost
- $20.78
- Avg. prompt tokens
- 1,664.9
- Avg. answer tokens
- 311.5
- Avg. reasoning tokens
- 0.2
- 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 | 92.6% |
| business | 89.5% |
| chemistry | 89.2% |
| computer science | 86.6% |
| economics | 88.7% |
| engineering | 81.2% |
| health | 80.1% |
| history | 80.8% |
| law | 73.2% |
| math | 94.1% |
| other | 82.6% |
| philosophy | 82.4% |
| physics | 90.3% |
| psychology | 87.8% |
Provenance
- Canonical model
- google/gemini-3-flash-preview
- Run ID
2026-06-25T16-07-36__vertex_ai__gemini-3-flash-preview__mmlu_prox_de