Individual experiment result
MMLU-ProX (DE)
Gemini 4 Argon Google
Exact match reasoning on / locked provider-internal
91.6%
Exact match
Run details
- Date
- 2026-10-13
- Test cases
- 11,759
- Median latency
- 4.28s
- Total cost
- $90.17
- Avg. prompt tokens
- 1,680.9
- Avg. answer tokens
- 299.1
- Avg. reasoning tokens
- 132.2
- 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 | 96.0% |
| business | 94.0% |
| chemistry | 92.3% |
| computer science | 93.2% |
| economics | 93.6% |
| engineering | 89.5% |
| health | 86.5% |
| history | 84.0% |
| law | 83.2% |
| math | 96.7% |
| other | 88.4% |
| philosophy | 89.2% |
| physics | 95.9% |
| psychology | 91.6% |
Provenance
- Canonical model
- google/gemini-4-argon
- Run ID
2026-10-13T19-23-38__vertex_ai__gemini-4-argon__mmlu_prox_de