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

Verifying agent behavior on AJ-Bench (2026): the judge decides whether each of 516 agent trajectories from search, data-system and GUI environments succeeded, and is scored against binary human-verified labels. In this setting the judge reads the trajectory only, with no tools or environment access. Mean F1 over the seven subdomains in percent averaged over three runs; higher is better.

Metric: Overall F1 (%), mean of the seven subdomain F1 scores and 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 February 2027. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Pro (Preview)75.05#119
2Claude Sonnet 4.569.77#261
3Grok 469.24#319
4Gemini 2.5 Pro68.31#276
5Claude Opus 4.566.79#160
6GLM-4.664.81#498
7DeepSeek V3.264.49#419
8Kimi K2 090564.33#570
9GPT-561.02#178
10GPT-5 MiniLow59#364
11Qwen 3 235B A22B58.47#628
12LongCat-Flash-Chat57.5#818
13GPT-5.153.5#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 · How It Works · Data refreshed daily, snapshot 2026-11-21.