Individual experiment result
MMLU-ProX (DE)
DeepSeek V4 Pro DeepSeek
Exact match reasoning off provider-internal
80.8%
Exact match
Run details
- Date
- 2026-06-16
- Test cases
- 11,759
- Median latency
- 3.00s
- Total cost
- $3.71
- Avg. prompt tokens
- 288.1
- Avg. answer tokens
- 218.9
- 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 | 90.2% |
| business | 85.3% |
| chemistry | 86.7% |
| computer science | 85.4% |
| economics | 84.1% |
| engineering | 74.2% |
| health | 74.2% |
| history | 71.9% |
| law | 55.3% |
| math | 90.9% |
| other | 77.3% |
| philosophy | 73.7% |
| physics | 86.8% |
| psychology | 83.3% |
Provenance
- Canonical model
- deepseek/deepseek-v4-pro
- Run ID
2026-06-16T13-49-10__openrouter__deepseek__deepseek-v4-pro__mmlu_prox_de