AgentViSS - Role Enactment: leaderboard

Metric: Mean of the expression-style and conflict-characteristic task scores under both observation modes; normalized task score (0-100), each role-task instance labeled Achieved 2, Partially Achieved 0.5 or Not Achieved 0 by majority vote of three judges (Gemini 3.1 Pro Preview, GPT-5.4, Qwen3.5-27B), over 2,340 role-task instances in 240 multi-party social scenarios with group images and role portraits; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 7 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.697.78
2GLM-4.6V97.63
3Qwen 3.5 27B96.81
4Qwen 3.5 122B A10B96.56
5Qwen 3.5 9B92.55
6GPT-5.486.37

Interactive version: theaggregate.ai/benchmark?slug=agentviss-role-enactment · How It Works · Data refreshed daily, snapshot 2026-09-29.