Atlas

Benchmarks

Concept Synth INDUCTION v1

INDUCTION v1 tests synthesis of one first-order concept formula from finite relational worlds. Its 775 solver-checkable instances span fully observed exact induction, contrastive induction, and existential completion under partial observation.

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

No signal or difficulty estimate. No usable scores.

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
Concept Synth INDUCTION v1 - CI13GPT-5.282.5higher is better2026-05-23
Concept Synth INDUCTION v1 - EC13GPT-5.494.0higher is better2026-05-23
Concept Synth INDUCTION v1 - FullObs13Grok 450.7higher is better2026-05-23
Loading Atlas data…