81.4%
Accuracy
Run details
- Date
- 2026-10-05
- Test cases
- 553
- Median latency
- 3.59s
- Total cost
- —
- Avg. prompt tokens
- 1,275.3
- Avg. answer tokens
- 1,176.1
- Avg. reasoning tokens
- —
- Quantization
- fp8
About this benchmark
German MuSR v1.1 defect-filtered benchmark: 553 questions (11 truncated team-allocation translations excluded), generate_until cot+; primary metric accuracy.
Subject breakdown
| Subject | Score |
|---|---|
| murder mystery | 79.6% |
| object placements | 64.1% |
| team allocation | 87.9% |
Provenance
- Canonical model
- aleph-alpha/kolibri-1-fp8
- Run ID
2026-10-05T17-09-03__openai__Aleph-Alpha__Kolibri-1__musr_de