85.0%
Accuracy
Run details
- Date
- 2026-08-30
- Test cases
- 553
- Median latency
- 8.65s
- Total cost
- —
- Avg. prompt tokens
- 1,889.7
- Avg. answer tokens
- 896.7
- Avg. reasoning tokens
- —
- Quantization
- Unverified
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 | 84.8% |
| object placements | 76.6% |
| team allocation | 87.4% |
Provenance
- Canonical model
- tencent/hy3
- Run ID
2026-08-30T19-46-08__openrouter__tencent__hy3__musr_de