Interactive AI Agent Benchmark & Model Matrix

Filter and sort models by SWE-bench Verified coding scores, Terminal-Bench v2 interactive execution scores, token costs, and license types.