ImmersedPrivacy - Social Context Selection (Native Audio): leaderboard

Metric: Selection accuracy (%; Tier 2: share of trials in which the model picks the candidate action with the highest mean human appropriateness rating for the observed social state; social-state audio or dialogue given as raw audio). Source: arxiv.org. Saturation forecast: Around November 2027. 6 models tracked.

Top models

#ModelScore
1Gemini 3 Flash (Preview) (High)54
2Gemini 3 Flash (Preview) (Low)49
3Gemini 3.1 Pro (Preview) (High)45
4Gemini 3.1 Pro (Preview) (Low)40

Interactive version: theaggregate.ai/benchmark?slug=immersedprivacy-social-context-selection-native-audio · How It Works · Data refreshed daily, snapshot 2026-09-26.