ICE-Guard - Authority: leaderboard

Share of decision pairs that flip when only a credential or prestige marker is swapped (for example a Harvard study versus a community college study) with identical evidence. Unlike the demographic slice, it measures deference to status rather than identity.

Metric: Decision flip rate under an irrelevant swap (%, lower is better). Source: arxiv.org. 10 models tracked.

Top models

#ModelScoreOverall rank
1GPT-OSS-120B1#700
2Claude Opus 4.62#102
3GLM-53.4#260
4Gemini 3 Flash4#164
5Gemini 2.5 Pro6#266
6GPT-5.26#176
7Kimi K2.56.3#256
8DeepSeek V3.27#406
9Qwen 3 32B8#960
10Claude Sonnet 4.510#250

Interactive version: theaggregate.ai/benchmark?slug=ice-guard-authority · How It Works · Data refreshed daily, snapshot 2026-11-13.