RevengeBench - Poker: leaderboard

Metric: Distance reduction (%; best round, mini-SWE-agent). Source: revengebench.com. Saturation forecast: Around August 2027. 12 models tracked.

Top models

#ModelScore
1GPT-5.582.2
2GLM-5.178
3GPT-576.3
4Qwen 3.6 35B A3B74.4
5DeepSeek V4 Flash73.6
6DeepSeek V4 Pro73
7Gemma 4 31B68.5
8Gemma 4 26B A4B63.6
9Grok 4.1 Fast60.6
10Kimi K2.658.4
11GPT-5.4 Mini44.2
12GPT-OSS-120B40.5

Interactive version: theaggregate.ai/benchmark?slug=revengebench-poker · How It Works · Data refreshed daily, snapshot 2026-10-02.