CyberGym
Progress Over Time
Interactive timeline showing model performance evolution on CyberGym
CyberGym Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Google | — | — | — | ||
| 2 |  Anthropic | — | — | — | ||
| 3 |  OpenAI | — | 1.1M | $5.00 / $30.00 | ||
| 4 | Anthropic | — | 1.0M | $5.00 / $25.00 | ||
| 5 | Anthropic | — | 1.0M | $5.00 / $25.00 | ||
| 6 | Anthropic | — | 1.0M | $5.00 / $25.00 | ||
| 7 |  ByteDance | — | — | — | ||
| 8 |  Zhipu AI | 754B | 200K | $1.40 / $4.40 | ||
| 9 | ByteDance | — | — | — | ||
| 10 |  Moonshot AI | 1.0T | — | — |
What is CyberGym?
CyberGym is a benchmark for evaluating AI agents on cybersecurity tasks, testing their ability to identify vulnerabilities, perform security analysis, and complete security-related challenges in a controlled environment.
CyberGym is a text benchmark evaluating models on safety, agents, and code tasks. LLM Stats tracks 10 models on this benchmark, scored on a 0–1 scale. The current average is 0.7, with the leader at 0.8.
Compare leaders on the best AI for safety, best AI for agents and best AI for code leaderboards.
Current leaders
Gemini 3.5 Flash Cyber from Google currently leads the CyberGym leaderboard with a score of 0.832 across 10 evaluated AI models.
FAQ
Common questions about the CyberGym benchmark and leaderboard.