● Covered by 2 sources · 2 reportsHigh impact2 neutral
OpenAI's Astra AI Model Solves 10 Long-Standing Math and Computer Science Problems
Key points
- OpenAI unveiled Astra, an unreleased AI model.
- Astra solved 10 long-standing math and computer science problems.
- Problems included high-dimensional geometry and coding theory.
- One problem was a counterexample to the unit distance problem by Paul Erdős.
- The solutions cost approximately $2,000 in API tokens.
OpenAI Introduces Astra
OpenAI has announced Astra, an unreleased AI model designed to address complex, long-running tasks. This announcement follows an internal version of the model making ten significant advances in mathematics and theoretical computer science. These problems had remained unsolved for at least a decade, and often much longer.
Significant Mathematical Breakthroughs
The internal research conducted by OpenAI covered a wide array of fields, including high-dimensional geometry, coding theory, arithmetic circuit complexity, group theory, quantum complexity, lattice cryptography, and extremal combinatorics. Specific examples of problems solved include the existence of non-sofic groups, a disproof of Connes’s rigidity conjecture, new bounds for high-dimensional sphere packing, and several problems posed by mathematician Paul Erdős.
One notable achievement, announced on May 20, 2026, involved an internal AI model generating a counterexample to the "unit distance" problem, a conjecture made by Paul Erdős in 1946. This marked the first historically significant proof from an AI model, introducing new ideas from a distant branch of mathematics. On August 1, Astra further contributed by solving three additional problems posed by Erdős.
Cost and Methodology
OpenAI noted that the total number of tokens required to find solutions to these problems would cost approximately $2,000 at Sol API rates. Human researchers utilized the same model to prepare the arguments as manuscripts, and Astra subsequently formalized each argument as a Lean certificate.
Implications for Scientific Research
These developments are seen by many mathematicians as a phase transition in the mathematical capabilities of AI models. The ability of AI to tackle and solve long-standing problems in foundational scientific fields suggests a changing landscape for mathematical research, potentially introducing new approaches to complex conjectures.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Primary sources
arXiv 2601.07421 arXiv 2605.00301 arXiv 2601.22401 arXiv 2605.22763 arXiv 2605.28781How outlets covered it
OpenAI's internal AI models have generated a counterexample to the unit distance problem and solved three other problems posed by mathematician Paul Erdős. These developments mark a significant shift in how mathematical research is conducted, introducing new approaches to long-standing conjectures.
OpenAI introduced Astra, an unreleased AI model designed for complex, long-running tasks, after an internal version made significant advances in mathematics and theoretical computer science. This development indicates a potential leap in AI's capability to solve foundational scientific problems, which could impact various research fields.