Skip to main content
MANIFOLD
High score on ExploitGym >= 70% by EOY 2026?
0
Ṁ100
Dec 31
56%
chance

GPT-5.6-Cyber @ 32.1% /as of August 2026, but there are also concerns that certain tasks in the eval suite are impossible.

Resolves to https://benchlm.ai/benchmarks/exploitgym and alternate sources from model providers if this website is not updated.

ExploitGym Leaderboard & Scores — August 2026
ExploitGym working exploit generation snapshot across 7 AI models. Display only on BenchLM and excluded from overall rankings. A controlled benchmark for evaluating whether AI agents can extend vulnerability-triggering inputs into working exploits.

(Identical market to: source market)

Market context
Mirror
AI Benchmarks
Get
Ṁ1,000
to start trading!

People are also trading