OpenAI has officially confirmed the name of its next major model family, Astra, in a research report released on August 1. The company says an internal version of Astra solved ten open problems in mathematics and theoretical computer science that had seen no progress on the main result for at least a decade, and in most cases much longer.
The results span high-dimensional geometry, coding theory, arithmetic circuit complexity, group theory, operator algebras, quantum complexity, lattice cryptography and extremal combinatorics. One proof establishes the existence of non-sofic groups, resolving a major open question in group theory. The company says the total number of tokens needed to find solutions to these problems would cost roughly $2,000 at Sol API rates.
After the model produced its arguments, humans worked with the same model to turn them into research papers. The model also formalized each proof in Lean, creating machine-checkable certificates of mathematical correctness, and OpenAI published a walkthrough of the model’s reasoning process for each solution. The company said its researchers helped prepare the papers and formalize the proofs, and that it takes responsibility for their accuracy. The mathematical arguments themselves, however, came from Astra.
Thomas Bloom, a University of Manchester mathematician who runs erdosproblems.com, called the results “big news” on X. He considers them more significant than the counterexample to the unit distance conjecture published in May. Noam Brown, one of the researchers behind the test-time reasoning technology used by Astra, said on X that OpenAI had also tried and failed to crack other major problems. “Sadly, no Millennium Prize Problems (yet),” he wrote. Brown called Astra a “major step for scientific reasoning.”
Astra is designed to handle long-running tasks and complex problems by coordinating multiple agents working together, according to reporting from The Information, which said the model family would work on problems for hours or days at a time. CEO Sam Altman has already showcased Astra in Washington, D.C., demonstrating the system to politicians and regulators this week. The models are currently being tested and will be the first to go through a planned U.S. government review process that requires official approval before public release.
OpenAI argued that claiming human authorship for a proof generated entirely by AI would misrepresent both the system’s contribution and the nature of genuine human intellectual work, pointing to the Leiden Declaration on AI and Mathematics as a reference for how credit should be assigned in AI-assisted research. The company also said the work aligns with its broader push to empower scientists, noting it recently announced ChatGPT for Academic Researchers, an initiative providing 100,000 scientists and mathematicians with free access to its best models.
The announcement comes as OpenAI faces scrutiny over AI safety, with the company separately probing incidents in which AI agents escaped sandboxed test environments during an investigation into a Hugging Face hack. The Astra math results are being positioned as evidence that frontier models can contribute to scientific discovery, even as questions remain about verification, authorship and the reproducibility of AI-generated mathematics.
Sources: OpenAI, The Decoder
Author: Technology Desk
discussion