InteractiveBench Trust Game: leaderboard
Iterated Prisoner's Dilemma trust-game domain with random horizon.
Metric: Average Payoff per Round (-1 to 3). Source: www.interactivebench.org. Source: www.interactivebench.org. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 Max | 1.87 | #208 |
| 2 | GPT-5 Mini | 1.84 | #186 |
| 3 | Grok 4.1 Fast | 1.8 | #217 |
| 4 | Kimi K2 (Thinking) | 1.78 | #243 (Kimi K2) |
| 5 | Gemini 3 Flash | 1.73 | #99 |
| 6 | DeepSeek V3.2 | 1.65 | #205 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=interactivebench-trust-game · How It Works · Data refreshed daily, snapshot 2026-10-20.