List your company

Theorem Proving on CLEVER human_eval

27Solved Tests (Correct)Numina Lean Agent
-1.086.2113.520.79May 22, 2026

Evaluation results

MethodLinks
Numina Lean AgentBenchmark variant=P2026.052714
Claude Code + lean4-skillsBenchmark variant=P2026.05268
AristotleBenchmark variant=PM2026.05239
Kimina-ProverBenchmark variant=P2026.053—
GrindBenchmark variant=PU2026.0531
LeanHammerBenchmark variant=PU2026.053—
AesopBenchmark variant=PU2026.052—
CanonicalBenchmark variant=PU2026.050—