CyberGym: Evaluating AI Agents' Cybersecurity Capabilities with Real-World Vulnerabilities at Scale

Record

Authors
Wang, Z.; et al.
Year
2025
Title
CyberGym: Evaluating AI Agents' Cybersecurity Capabilities with Real-World Vulnerabilities at Scale
Published in
arXiv
Type
Preprint
arXiv
arXiv:2506.02548
Our note
none yet
Verification
verified: its identifier resolves to a record with the same title and author (checked against arXiv, 2026-11-14).
Identifiers found: arXiv:2506.02548
Notes: title of v1; the current version is titled "CyberGym: Evaluating AI Agents' Real-World Cybersecurity Capabilities at Scale".
Id
wang-2025-cybergym-evaluating-ai-agents-cybersecurity-capabilities

Cited on

As cited

The pages cite this work in one form.

BibLaTeX

@misc{wang-2025-cybergym-evaluating-ai-agents-cybersecurity-capabilities,
  author = {Wang, Z. and others},
  title = {{CyberGym: Evaluating AI Agents' Cybersecurity Capabilities with Real-World Vulnerabilities at Scale}},
  date = {2025},
  eprint = {2506.02548},
  eprinttype = {arxiv}
}

Download · JSON with CSL-JSON · the whole bibliography