Individual experiment result
MMLU-ProX (DE)
Claude Haiku 4.5 Anthropic
Exact match reasoning off provider-internal
75.3%
Exact match
Run details
- Date
- 2026-06-03
- Test cases
- 11,759
- Median latency
- 3.76s
- Total cost
- $52.10
- Avg. prompt tokens
- 2,262
- Avg. answer tokens
- 433.7
- 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 | 85.9% |
| business | 79.0% |
| chemistry | 80.4% |
| computer science | 83.4% |
| economics | 81.5% |
| engineering | 64.7% |
| health | 72.9% |
| history | 65.4% |
| law | 47.4% |
| math | 86.8% |
| other | 69.5% |
| philosophy | 67.9% |
| physics | 81.3% |
| psychology | 78.9% |
Provenance
- Canonical model
- anthropic/claude-haiku-4.5
- Run ID
2026-06-03T15-52-51__openai__cc__claude-haiku-4-5-20251001__mmlu_prox_de