OpenAI has launched GPT-6 Astra, the most intelligent and aligned model OpenAI has ever released.
Astra represents a generational leap in capability, delivering state-of-the-art performance on computer use, browser use, software engineering, cybersecurity, science, and professional work. GPT-6 Astra is built on advances across pre-training, reinforcement learning, and alignment, bringing together years of research and big bets.
A few key things to know:
- Fast, reliable, and accurate computer use: Astra marks a new frontier in the speed, accuracy, and safety of computer use. It can carry out multistep workflows, produce polished documents, spreadsheets, and presentations, and create websites and test that their features work. In latency simulations on the offline subset of OSWorld 2.0, Astra achieves higher computer-use performance in about 47% less time per task than GPT-5.6 Sol.
- Advancing scientific discovery: Astra is already contributing to new mathematical knowledge. Last month, we shared ten advances in mathematics and theoretical computer science produced by an internal version of Astra, with proofs formalized in Lean. Astra also scores 98% on FrontierMath Tier 4.
- Helping developers tackle harder work more efficiently: Astra is our best model for software engineering, with stronger performance on complex tasks in real codebases. On DeepSWE v1.1, it outperforms GPT-5.6 Sol at approximately 57% lower estimated API cost per task, comparing each model’s highest-scoring setting. An experimental, opt-in Codex feature also helps Astra stay oriented during long projects by preserving notes and retrieving earlier requirements, attempted fixes, and test results across context windows.
- Greater confidence in the work you delegate: Astra is our most aligned model yet, with substantial improvements in understanding user intent, exercising care, respecting task boundaries, and communicating transparently. It uses context to fill in routine gaps, asks focused questions when the answer could change the outcome, and waits for input on consequential decisions—helping people delegate complex work while maintaining oversight.
- Helping cybersecurity defenders keep pace: Astra’s stronger cyber capabilities can help defenders find and patch weaknesses, while also creating a need for stronger safeguards. It meets the Critical threshold in cybersecurity under our Preparedness Framework. We’re strengthening protections against misuse. Through OpenAI Daybreak, we plan to expand access and roll out less restrictive safeguards in the coming weeks, supporting work such as vulnerability validation, malware analysis, and detection engineering.
GPT-6 Astra is rolling out today to a limited set of organizations and over the coming days will become available to all ChatGPT Plus, Pro, Business, and Enterprise users, as well as through the OpenAI API and AWS.
If you have an interesting Article / Report/case study to share, please get in touch with us at editors@roymediative.com roy@roymediative.com, 9811346846/9625243429
