Blog post · 2026

Demystifying evals for AI agents

Anthropic

URLhttps://www.anthropic.com/engineering/demystifying-evals-for-ai-agents
TopicsReplicHAI / Agent Evals / 10 Harnesses, leaderboards & methodology
Tagsagent-eval

Abstract

Measures: Practical guide: tasks, graders, transcripts, pass@k vs pass^k. Metric: -.

BibTeX

@misc{anthropic_demystifying_2026,
	title = {Demystifying evals for {AI} agents},
	url = {https://www.anthropic.com/engineering/demystifying-evals-for-ai-agents},
	abstract = {Measures: Practical guide: tasks, graders, transcripts, pass@k vs pass{\textasciicircum}k. Metric: -.},
	author = {{Anthropic}},
	month = jan,
	year = {2026},
	keywords = {agent-eval},
}