Webpage · 2026

Terminal-Bench 4.0 leaderboard

Snorkel AI / Laude Institute

URLhttps://snorkel.ai/leaderboard/terminal-bench/
TopicsReplicHAI / Agent Evals / 02 Software engineering & coding
Tagsagent-eval, gap-sweep-2026-11

Abstract

Measures: Continuously updated terminal-agent software tasks. Metric: % tasks passed.

BibTeX

@misc{snorkel_ai__laude_institute_terminal-bench_2026,
	title = {Terminal-{Bench} 4.0 leaderboard},
	url = {https://snorkel.ai/leaderboard/terminal-bench/},
	abstract = {Measures: Continuously updated terminal-agent software tasks. Metric: \% tasks passed.},
	author = {{Snorkel AI / Laude Institute}},
	year = {2026},
	keywords = {agent-eval, gap-sweep-2026-11},
}