OpenAI launches Astra for enterprise work

OpenAI has introduced Astra as its most capable model for business work, making it available in ChatGPT Work, Codex, and the API. The model is designed to operate across everyday applications, including those without APIs, allowing teams to use AI inside existing workflows for tasks such as software engineering, financial analysis, slide creation, browsing, cybersecurity, and scientific work.

Early users and internal teams reported gains in complex professional tasks. OpenAI said its teams used Astra and Codex to turn three hours of multicamera footage into a developer video that drew over 550k views in just 4 days, while engineers resolved a memory-allocation bottleneck that produced 25× lower turn latency with roughly 30% higher peak memory use. Pricing starts at $10 per million input tokens and $50 per million output tokens.

OpenAI also emphasized safety and administrative control. On an internal computer use safety benchmark, Astra produced unintended outcomes 89% less often than GPT‑5.6 Sol and 74.7% less often than Claude Fable 5.1. Enterprise admins can restrict approved websites and desktop apps, manage uploads and downloads, control browsing history, and require confirmations before consequential actions.