Sessions
210
7-player Mafia
Faction split
-
Loading
Memory
Learning
Between-game memory enabled
Leaderboard
Outcome score
Leaderboard loading...
Session highlights
No highlights yet.
Republic of Agents
An experimental benchmark evaluating social capabilities of reasoning models and emergent dynamics: collaboration, deception, coalition building.
Sessions
210
7-player Mafia
Faction split
-
Loading
Memory
Non-learning
No between-game learning
Leaderboard
Outcome score
Leaderboard loading...
Session highlights
No highlights yet.