OpenAI Is Building Astra: AI That Can Work for Hours and Days at a Time

OpenAI is developing a new model family codenamed Astra — an AI capable of working on a single task for hours or even days without stopping. According to The Information (three independent sources), the system simultaneously coordinates multiple agents — individual AI "employees" — and is aimed at the most demanding challenges: math, science, and multi-step long-term research.

Sam Altman already previewed Astra to policymakers in Washington this week. The system is going through a new vetting program run by the U.S. administration — the first OpenAI model to enter that process. The name and release date haven't been finalized: Astra may launch under the GPT-6 or GPT 5.7 brand.

The context is bigger than a single product. OpenAI has already announced plans to build an autonomous AI researcher by March 2028 and a research-intern-level system by September 2026. Astra is a step in that direction: not one powerful chatbot, but a coordinated swarm of specialized agents.

This is a fundamental shift in what counts as normal. AI is moving from an ask-and-answer tool to a system that carries a project from start to finish. If Astra delivers on its promise, the line between AI assistant and AI employee will disappear entirely — and the tech labor market will be the first to feel it.