Individual experiment result
MuSR (DE)
Gemini 3.5 Flash-Lite Google
Accuracy reasoning off provider-internal
82.3%
Accuracy
Run details
- Date
- 2026-08-30
- Test cases
- 553
- Median latency
- 3.13s
- Total cost
- $1.15
- Avg. prompt tokens
- 1,484.9
- Avg. answer tokens
- 653
- Avg. reasoning tokens
- —
- 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 | 85.2% |
| object placements | 75.0% |
| team allocation | 81.2% |
Provenance
- Canonical model
- google/gemini-3.5-flash-lite
- Run ID
2026-08-30T07-13-28__vertex_ai__gemini-3.5-flash-lite__musr_de