AJ-Bench (LLM-as-a-Judge) - Wide Search: leaderboard

Verifying agent trajectories on AJ-Bench (2026), the Wide Search subdomain (27 trajectories of broad information-collection tasks): the judge decides whether each trajectory succeeded and is scored against binary human-verified labels. In this setting the judge reads the trajectory only, with no tools or environment access. F1 in percent; higher is better.

Metric: F1 (%) on Wide search (WideSearch-derived item checks; 9 tasks, 27 trajectories), averaged over three runs, with the judge reading the trajectory only (no tools or environment access), of AJ-Bench, judging whether agent trajectories succeeded against binary human-verified labels (155 tasks, 516 trajectories), default model configurations; higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Pro (Preview)72.7#119
2Grok 469.18#319
3GLM-4.666.61#498
4Gemini 2.5 Pro66.35#276
5GPT-566.33#178
6LongCat-Flash-Chat64.44#818
7Claude Opus 4.564.26#160
8DeepSeek V3.263.65#419
9Kimi K2 090563.52#570
10Qwen 3 235B A22B62.69#628
11Claude Sonnet 4.561.02#261
12GPT-5 MiniLow60.84#364
13GPT-5.158.02#240

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=aj-bench-llm-as-a-judge-wide-search · How It Works · Data refreshed daily, snapshot 2026-11-21.