The model accepts text and images and returns text. Applications can give it investigation context and use its reasoning in a controlled workflow, with access managed through OpenAI's Daybreak Red program and organizational security requirements.
Approved security teams can use the model to work through larger collections of evidence and document results for review. Its long context is designed for complex investigations that involve multiple files, findings and iterations.
