Individual experiment result
MuSR (DE)
Gemini 3.7 Flash Google
Accuracy reasoning on / locked provider-internal
88.6%
Accuracy
Run details
- Date
- 2026-08-14
- Test cases
- 553
- Median latency
- 8.23s
- Total cost
- $2.37
- Avg. prompt tokens
- 1,484.9
- Avg. answer tokens
- 372.5
- Avg. reasoning tokens
- 474.1
- 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 | 91.2% |
| object placements | 75.0% |
| team allocation | 89.5% |
Provenance
- Canonical model
- google/gemini-3.7-flash
- Run ID
2026-08-14T13-26-58__vertex_ai__gemini-3.7-flash__musr_de