Technology
Add as a preferred source on GoogleGPT-5.6 Cyber scored 32.1% in the same comparison.
OpenAI's ExploitGym comparison returned scores of 48.5% for GPT-6 Cyber and 32.1% for GPT-5.6 Cyber, a 16.4-percentage-point spread.
ExploitGym's development drew on academic research and feedback and practical support from industry partners.
EconoTimes previously included the scores in its coverage of GPT-6 Cyber's Linux exploit results.
OpenAI also said the model reproduced 1,724 Linux kernel vulnerabilities and developed 948 working privilege-escalation exploits.
The planned system is intended to reduce manual work across cybersecurity tasks.
The broader agreement includes engineering and other bank operations. JPMorgan’s LLM Suite had about 150,000 active users each week by early 2026.
