OpenAI said on November 5, 2026, that GPT-6 Cyber produced evidence and reproducible tests for 81 previously unknown Firefox vulnerabilities that Mozilla classified as high severity. OpenAI said it sent the findings to Mozilla with proposed patches and reported that Mozilla had released fixes for many of them. OpenAI’s announcement describes the findings and remediation.
Reproducible tests and proposed patches connect a security finding to the engineering work of fixing it.
What OpenAI reported about Firefox
OpenAI’s report says GPT-6 Cyber produced evidence and tests that could reproduce 81 new Firefox vulnerabilities. Mozilla classified those findings as high severity, according to OpenAI. The company also said the model identified 30 previously known issues that were more severe than earlier assessments had indicated.
How the findings fed into Mozilla’s fixes
Brian Grinstead, a Mozilla Distinguished Engineer for Firefox, said reproducible test cases and fix patches helped engineers land fixes quickly. A reproducible test lets engineers trigger the same bug again, giving them a concrete case to investigate and address.
