JPMorgan plans to use OpenAI’s Codex and cybersecurity models for internal security work, while the broader agreement also reaches engineering and other operations at the bank.
OpenAI President Greg Brockman announced the arrangement Nov. 4. He said trust and regulatory standards would guide the work.
The bank began using OpenAI models in its LLM Suite in 2024. Hundreds of thousands of employees can use the suite, which also includes models from Anthropic. By early 2026, the platform had about 150,000 active users per week.
Brockman said enterprise now generates most of OpenAI’s revenue, compared with about 30% at an earlier point. EconoTimes previously covered OpenAI’s enterprise sales push.
