Benchmarks
347 sources · 1,258 models × 3,038 benchmarks · 158,912 scoresSOLO Bench asks a model to create unique four-word sentences under strict grammar and vocabulary constraints without tools. It measures long-output instruction following, memory, and hallucination resistance with deterministic rule-based scoring.
No signal or difficulty estimate. Held out of the fit by design.
This is a benchmark-family page. Scores are recorded on the component benchmarks below rather than directly on the parent, because the components use different score scales and directions.
| Metric | Models | Leader | Direction | Date |
|---|---|---|---|---|
| SOLO Bench Easy | 21 | Gemini 2.5 Pro Preview 03-2574.8 | higher is better | 2025-05-01 |
| SOLO Bench Medium | 6 | Gemini 2.5 Pro Preview 03-2557.8 | higher is better | 2025-05-01 |