OpenAI is preparing to launch a new model series named Astra, which focuses on multi-agent long-term task collaboration.

According to three people familiar with the matter, OpenAI is preparing to launch a new model series tentatively named "Astra", which will feature enhanced capabilities for completing long-cycle tasks. OpenAI founder Sam Altman presented Astra to policymakers and regulators in Washington this week. Reportedly, OpenAI touts that the model enables multiple AI agents to collaborate over extended periods to solve highly complex problems, with applications in projects or advanced mathematical tasks. Astra will be a new category of models for OpenAI, following Sol, Terra, and Luna. It remains unclear when OpenAI plans to release the model. The company has also not yet decided whether to brand Astra as GPT-6 or an add-on to the GPT-5 series, such as GPT-5.7. According to one insider, while the models are already in testing, they are expected to be among the first AI models submitted under a new framework planned by the Trump administration, which requires AI models to be filed with the federal government before public release. The administration has set a self-imposed deadline of the end of this week, aiming to finalize the framework by then. Reportedly, OpenAI also plans to release a report in the near term detailing how it solved 10 previously unsolved mathematical problems, to showcase the capabilities of its most advanced AI. (The Information)