Atlas

Benchmarks

SOLO Bench

SOLO 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.

Models
0
Top score
—
Leader
—
Updated
—
Resolution
—
Saturation
—

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.

MetricModelsLeaderDirectionDate
SOLO Bench Easy21Gemini 2.5 Pro Preview 03-2574.8higher is better2025-05-01
SOLO Bench Medium6Gemini 2.5 Pro Preview 03-2557.8higher is better2025-05-01
Loading Atlas data…