NormGym - Joint Success (Seen Layout): leaderboard

Interactive social-norm learning benchmark (arXiv:2611.08613): share of 150 seen-layout test episodes across 50 hidden norms where a ReAct LLM/VLM agent completes both task goals with zero norm violations.

Metric: Joint Task & Norm-Compliance Success Rate (%), Seen Layout. Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked. Updated 20 Nov 2026.

Top models

#ModelScoreOverall rank
1Claude Opus 555.3#26
2GPT-5.6 Sol44#36
3Gemini 3.1 Pro (Preview)24#101
4Ministral-3-8B-Instruct3.3#1299
5Qwen 3 VL 8B Instruct1.3#1035
6Llama 3.2 11B Instruct0#5189
7GLM-4.6V-Flash0#1047

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

Interactive version: theaggregate.ai/benchmark?slug=normgym-joint-success-seen-layout · How It Works · Data refreshed daily, snapshot 2026-11-20.