AgentEscapeBench - Difficulty 5: leaderboard

Metric: Success rate (%; share of the 60 difficulty-5 instances (dependency DAGs of 5 tool and item nodes) whose deterministic final flag the agent submits; single run per instance, official APIs at temperature 0.7, multi-turn tool calling with incremental disclosure of hidden nodes). Source: arxiv.org. Saturation forecast: Estimated already saturated. 16 models tracked.

Top models

#ModelScore
1GPT-5.496.7
2Kimi K2.595
3Gemini 3.1 Pro (Preview)91.7
4Claude Opus 4.690
5Gemini 3 Flash (Preview)88.3
6Qwen 3 235B A22B88.3
7Seed 2.0 Pro86.7
8Qwen 3 Next 80B A3B76.7
9MiniMax-M270
10Qwen 3 32B60
11GPT-556.7
12Qwen 3 14B50
13Qwen 3 8B28.3

Interactive version: theaggregate.ai/benchmark?slug=agentescapebench-difficulty-5 · How It Works · Data refreshed daily, snapshot 2026-09-26.