On September 8, 2026, OpenAI released an AI-generated solution to the Navier-Stokes Millennium Prize Problem, accompanied by a detailed writeup and a machine-checked formal proof in the Lean theorem prover, marking the first time the organization shared a complete machine-produced candidate for a Millennium Prize challenge.
Short answer: On September 8, 2026, OpenAI released an AI-generated solution to the Navier-Stokes Millennium Prize Problem, accompanied by a detailed writeup and a machine-checked formal proof in the Lean theorem prover, marking the first time the organization shared a complete machine-produced candidate for a Millennium Prize challenge.
On September 8, 2026, OpenAI announced that it is making available an AI-generated answer to the Navier-Stokes Millennium Prize Problem. The release includes a detailed writeup that explains the approach taken by the model and a formal proof of the result written in the Lean theorem prover. This marks the first time the organization has shared a complete machine-produced candidate for one of the seven celebrated Millennium Prize challenges.
The Navier-Stokes equations describe the motion of fluids and have resisted a rigorous solution for decades. By presenting an AI-derived answer, OpenAI is showing how modern language models can be directed toward deep mathematical reasoning. The accompanying writeup walks readers through the model’s internal steps, highlighting where it introduced new definitions, how it structured the argument, and where it relied on existing mathematical knowledge.
In addition to the narrative explanation, the team supplied a machine-checked proof in Lean. Lean is a proof assistant that requires every logical step to be justified according to a strict formal language. Including a Lean script means that anyone can verify the correctness of the argument automatically, reducing reliance on human review alone. This combination of a high-level exposition and a low-level formal artifact offers a clear pathway for other researchers to examine, build on, or challenge the work.
For people who build or use AI systems, the release is a concrete demonstration that large models can contribute to problems that have long been considered beyond the reach of automated methods. It suggests that pairing generative models with formal verification tools may open new avenues for tackling other unsolved questions in mathematics, physics, and computer science. Practitioners might consider how to integrate similar pipelines into their own workflows, using AI to generate conjectures or proof sketches and then employing proof assistants to validate them.
Readers interested in exploring the material should start by reading the writeup to grasp the overall strategy. After that, they can download the Lean file and run it through the Lean environment to see the proof check in real time. If any step appears unclear, the writeup provides pointers to the relevant lemmas and definitions that the model invoked. Engaging with both layers, conceptual and formal, helps assess the strength of the AI-generated argument and identifies areas where further human insight might be needed.
The announcement also raises questions about how the AI community evaluates success in mathematical discovery. Traditional metrics focus on novelty and correctness, but the process of generating a formal proof introduces additional considerations such as proof length, readability, and the degree of automation versus human guidance. Discussions around these topics are likely to grow as more groups experiment with AI-assisted theorem proving.
Finally, while the release is a notable step, it does not yet claim the official Millennium Prize, which requires a solution verified by the designated awarding body. Nonetheless, sharing the work openly invites scrutiny from the broader mathematical community and may accelerate progress toward a universally accepted resolution. By examining the writeup and the Lean proof, AI developers can gain practical insight into how to steer powerful models toward rigorous, verifiable outcomes in domains that demand absolute precision.
On September 8, 2026, OpenAI announced it is sharing an AI-generated answer to the Navier-Stokes Millennium Prize Problem, including a detailed writeup of the model’s approach and a machine-checked proof written in the Lean theorem prover.
Besides the narrative writeup, the release supplies a formal proof in Lean, a proof assistant that demands every logical step be justified in a strict language, allowing anyone to verify the argument automatically.
Researchers should first read the writeup to understand the overall strategy, then download the Lean file and run it in the Lean environment to see the proof check in real time, using the writeup’s pointers for any unclear steps.
No, the release does not yet claim the official Millennium Prize, which requires verification by the designated awarding body; it merely shares the work openly for community scrutiny and potential progress toward a universally accepted resolution.
Source: OpenAI
On September 8, 2026, OpenAI released ChatGPT Images 2.5 with a new Sketch feature that lets users type @Sketch, draw a rough doodle inside the chat, add a natural-language description, and receive an AI-generated image that blends the sketch with the instructions, allowing inline edits and up to 50 % faster generation.
On September 8, 2026, OpenAI released ChatGPT Images 2.5, adding a Sketch feature that lets users draw a rough doodle in a pop-up canvas via @Sketch, then describe style or details to generate a polished AI image.
OpenAI announced on September 8 2026 that its researchers solved the Navier-Stokes equation using an internal AI system more capable than GPT-6 Astra, aided by ten thousand simultaneous processing agents; the claim is contested by allegations that the model may have benefited from user data in Codex sessions.
NO COMMENTS YET
Comments are open. Have a thought or a question? Share it below.