Public KG record
GPT-6 Cyber
Recorded overview
A cybersecurity-focused model trained to find and validate exploitable vulnerabi
An evidence-backed entity from the project knowledge graph; detailed tool metadata is still being added.
Recorded details
- Benchmarks: ExploitGym 48.5%
- Benchmarks: Offensive web penetration testing 40.3% recall
- Benchmarks: RemoteExploitBench 95% completion rate
- Benchmarks: 48.5% on ExploitGym
- Benchmarks: 40.3% recall on web penetration testing evaluation
- Benchmarks: 99.93% success rate avoiding out-of-scope services without safety prompts
- Benchmarks: ExploitGym
- Benchmarks: RemoteExploitBench
Evidence coverage
The KG currently records 7 known information dimensions and has 10 undocumented dimensions, linked to 1 public source articles. These counts describe record coverage, not model performance or verified fit.
Related public records
- ExploitGym — evaluated on
- RemoteExploitBench — evaluated on
- Preparedness Framework — evaluated on
- GPT-6 Astra — outperforms
- Chain of Thought — uses