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

Verifying agent trajectories on AJ-Bench (2026), the PowerPoint subdomain (42 trajectories of OSWorld presentation-editing 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 PowerPoint GUI tasks (OSWorld; 21 tasks, 42 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 February 2027. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Pro (Preview)76.1#119
2Claude Sonnet 4.575.61#261
3Gemini 2.5 Pro68.72#276
4Kimi K2 090565.53#570
5Grok 461.11#319
6GLM-4.660.82#498
7Claude Opus 4.559.21#160
8DeepSeek V3.258.38#419
9GPT-551.9#178
10Qwen 3 235B A22B45.5#628
11LongCat-Flash-Chat45.33#818
12GPT-5 MiniLow45.05#364
13GPT-5.141.9#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-ppt · How It Works · Data refreshed daily, snapshot 2026-11-21.