Social Gym - Bargaining: leaderboard

Metric: Bradley-Terry Elo rating (regularized fit on the pooled cross-model pairwise outcomes of the Bargaining game (structured negotiation over a divisible payoff, the one competitive economic game), scale 400, anchored at a mean of 1000 over the seven evaluated models; seats and roles balanced over every model combination; relative to this model pool only). Source: arxiv.org. Saturation forecast: Around November 2026. 7 models tracked.

Top models

#ModelScore
1GPT-4o1131
2GPT-4o Mini1067
3Qwen 2.5 3B Instruct1060
4GPT-5 Mini1039
5Qwen 3 4B1031
6Qwen 3 32B837
7Gemma 3 27B831

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