Your security vendor gets the frontier cyber model, you get the findings

Selected red team specialists can now use OpenAI’s cyber models to find and exploit weaknesses in client applications and infrastructure. Those clients never...