AI Model Rankings

Overall AI model ranking across HLE, ARC-AGI-2, FrontierMath, SWE-bench Verified, and τ²-Bench.

5 benchmarks

Programming ability benchmarks: SWE-bench Verified, LiveCodeBench, SWE-bench Pro, Aider-Polyglot.

4 benchmarks

Mathematical reasoning benchmarks: AIME 2025/2026, FrontierMath, MATH-500, GSM8K.

5 benchmarks

Autonomous agent benchmarks: τ²-Bench, Terminal Bench Hard, Aider-Polyglot.

3 benchmarks

Reasoning and thinking benchmarks: HLE, ARC-AGI-2, GPQA Diamond.

3 benchmarks

General AI performance: MMLU-Pro, LMArena Elo ratings.

2 benchmarks

OpenClaw agent performance: Claw Bench and Pinch Bench.

2 benchmarks