Individual experiment result
MMLU-ProX (DE)
Gemini 3.1 Flash-Lite Google
Exact match reasoning off provider-internal
82.2%
Exact match
Run details
- Date
- 2026-06-03
- Test cases
- 11,759
- Median latency
- 1.24s
- Total cost
- $10.21
- Avg. prompt tokens
- 1,666.1
- Avg. answer tokens
- 304.5
- Avg. reasoning tokens
- —
- 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 | 89.4% |
| business | 85.3% |
| chemistry | 87.1% |
| computer science | 88.8% |
| economics | 87.3% |
| engineering | 77.8% |
| health | 74.8% |
| history | 74.5% |
| law | 62.8% |
| math | 90.2% |
| other | 76.7% |
| philosophy | 75.2% |
| physics | 87.6% |
| psychology | 83.3% |
Provenance
- Canonical model
- google/gemini-3.1-flash-lite
- Run ID
2026-06-03T16-05-15__gemini__gemini-3.1-flash-lite__mmlu_prox_de