GPT-6 Astra is OpenAI's latest general-purpose model, introduced as "a new generation of intelligence" that brings together years of research across pre-training, reinforcement learning, and alignment. The model is built for agentic work: driving computers, browsing the web, writing and executing code, operating software, working across large file collections, and continuing complex, multi-step tasks with less human guidance. It is also OpenAI's first model to cross the company's internal "Critical" cybersecurity capability threshold, which is why initial access was limited to organizations in its application-based Daybreak program before broader rollout through ChatGPT Plus, Pro, Business, and Enterprise, the OpenAI API, and Amazon Web Services.

In benchmark results published by OpenAI, GPT-6 Astra is reported as state-of-the-art across computer use, browsing, software engineering, cybersecurity, science, and professional work. The company says it saturates FrontierMath Tier 4 with a 98% score, ARC-AGI-3 with 99.9%, and ExploitBench with 100%, while Coursiv reports a 1.05-million-token context window and up to the cataloged API limit output tokens. Practically, that combination positions Astra for demanding professional workflows, from long-horizon research and software engineering to autonomous browsing and agent-style task execution, with strong results on terminal and science benchmarks that test whether models can complete real research and coding workflows rather than answer isolated questions.