← all datasets

Terminal-Bench-2

Emerging

4papers using it

2026first seen

🔎 Find this dataset

Papers using Terminal-Bench-2 (4)

Dissecting model behavior through agent trajectories2026

Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses2026 · 1 cites

Shepherd: A Runtime Substrate Empowering Meta-Agents with a Formalized Execution Trace2026

Automated Benchmark Auditing for AI Agents and Large Language Models2026

Terminal-Bench-2 — datasets — ai-agents