Zhun Wang et al.
ExploitGym
A controlled benchmark for evaluating whether AI agents can extend vulnerability-triggering inputs into working exploits.
ModelsModels with a result.23
Top-3 spreadPoints from first to third.11.1 pts
YearYear of release.2026
Result and price
21 of 23 results
Results
23 results#ModelResultThe score from the source.
