Individual experiment result
MuSR (DE)
Muse Spark Open Meta
Accuracy reasoning on / locked provider-internal
87.9%
Accuracy
Run details
- Date
- 2026-10-16
- Test cases
- 553
- Median latency
- 12.73s
- Total cost
- $5.26
- Avg. prompt tokens
- 1,588
- Avg. answer tokens
- 661.2
- Avg. reasoning tokens
- 1,138.3
- Quantization
- provider-internal
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 | 88.0% |
| object placements | 79.7% |
| team allocation | 90.0% |
Provenance
- Canonical model
- meta/muse-spark-open
- Run ID
2026-10-16T11-38-00__openrouter__meta__muse-spark-open__musr_de