Terminal-Bench

Terminal-Bench is an open-source agent evaluation framework developed by Harbor for benchmarking AI coding agents on real-world terminal tasks. It differentiates itself by simulating complete software engineering workflows to measure agent autonomy a

Two-hop subgraph: this entity, every entity it directly relates to, and every entity those neighbors relate to. Drag a node, scroll to zoom, click to inspect — or click any neighbor and re-center the atlas there.