AgentEvalBench (EvalAgent): leaderboard
Automated agent evaluation on AgentEvalBench (Zhou et al., 2026, AWS AI Labs): for 20 real LLM agents with pre-collected execution traces, each under a generic and a specific requirement, the system writes evaluation code for the agent through EvalAgent, the system the paper proposes: Claude Code with evaluation skills (procedures, reusable templates and Context7 API documentation) and a plan-then-implement pipeline. Eval@1 in percent over the 40 cases, counting code that runs and produces non-vacuous results on the real traces at the first attempt; code that runs on mocked data or returns constant metrics fails. Higher is better.
Metric: Eval@1 (%; share of 40 cases, 20 agents under generic and specific requirements, whose generated evaluation code runs and yields non-vacuous results on real agent traces at the first attempt; the authors' EvalAgent: Claude Code v2.1.92 with evaluation skills and Context7 documentation). Source: arxiv.org. Saturation forecast: Around February 2027. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 65 | #244 |
| 2 | Claude Haiku 4.5 | 62.5 | #568 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=agentevalbench-evalagent · How It Works · Data refreshed daily, snapshot 2026-11-12.