🌀 OpenAI Claims Navier-Stokes Proof

ChatGPT Images Gets Upgrade

Author

The Summary AI
September 09, 2026

In partnership with

Welcome back!

AI has cracked one of mathematics’ hardest problems. OpenAI says 10,000 AI agents produced a proof for the Navier-Stokes Millennium Problem. The result is historic, though it has also let to a dispute with two mathematicians who were already close to a solution. Let’s unpack…

Today’s Summary:

  • 🌀 OpenAI claims Navier-Stokes proof

  • 🖼️ OpenAI launches ChatGPT Images 2.5

  • 🧬 Google AI maps DNA mutations

  • đź’» Meta launches personal AI agent

  • 🛡️ OpenAI agents bypass safety limits

  • ⏸️ OpenAI scientist warns about slowdowns

  • 🛠️ 2 new tools

TOP STORY

OpenAI claims solution to Navier-Stokes math problem after research race

The Summary: OpenAI produced a proof for the Navier-Stokes Millennium Prize Problem, one of the seven hardest problems in mathematics, using 10,000 agents and several million dollars of compute. Days earlier, rumors said Anthropic’s Levent Alpöge and NYU professor Tristan Buckmaster were close to solving it, and this prompted OpenAI to start its own effort to solve the problem as well, leading to a dispute over credit, data use, and research practices. Behind the drama sits a remarkable fact: current AI is able to produce a proof for a major problem that remained unsolved for nearly 90 years.

Key details:

  • The proof shows that Navier-Stokes dynamics can develop a singularity in finite time, with a vortex winding inward and stretching like spaghetti

  • Alpöge and Buckmaster spent a year working on the problem with help from Claude and OpenAI models, and were already close to a solution

  • OpenAI says nobody directly accessed their chats, while also saying it cannot rule out that de-identified usage data helped improve its models

  • Buckmaster says OpenAI offered to let him present its proof while excluding Alpöge because he works at Anthropic

Why it matters: For nearly 90 years, no one could solve Navier-Stokes. AI found a proof in 4 days. One mind can test only a few paths at a time. Ten thousand agents can test more. This result may be an early glimpse of science moving from finding results to finding the right questions to ask, and understanding what the answers mean.

FROM OUR PARTNERS

What Happens When One AI Replaces 10 Tools

Stop Paying for 10 Tools. One AI Does It All.

Most e-commerce sellers are running their store across 6 to 10 separate tools — and spending more time managing software than growing their business. StoreClaw replaces your entire stack with one autonomous AI engine that monitors competitors, optimizes listings, automates marketing, and tracks real profit across Shopify, Amazon, and beyond.

It doesn't wait for you to ask. It runs 24/7 in the background, so you wake up to a full dashboard instead of a list of things you forgot to check.

Connect your store, and StoreClaw gets to work — no prompts, no complex setup, no six-app stack.

Free to start. No credit card required.

OPENAI

OpenAI launches ChatGPT Images 2.5

The Summary: OpenAI has released ChatGPT Images 2.5, bringing faster generation, stronger fidelity to reference images, and more precise edits. The model can keep subjects and layouts intact through long editing sessions, so repeated edits hold up better. ChatGPT also added a Sketch tool, templates, image comments, and prompt sharing.

Key details:

  • Users now create more than 3 billion images each week with ChatGPT Images and GPT-Image models

  • OpenAI released two new API models: GPT-Image-2.5 Flare runs about 2x faster, while Sunburst is built for work that needs tight edit control

  • Arena.ai ranked Sunburst #1 at 1421, 40 points above GPT-Image-2 in text-to-image and an 81-point gain in multi-image editing

Why it matters: AI images are easy to create but annoying to edit. Images 2.5 is improving the second part: making precise changes without ruining the rest of the image. It is also much faster. At around 30–40 seconds per attempt, more ideas can be tested in the same session.

FROM OUR PARTNERS

Go Global in Days

Hire anyone, anywhere — compliant in under 3 days

Found the right hire, but no entity in their country? Remote becomes the legal employer — with contracts, benefits, and tax setup handled, plus direct access to the same in-house team that runs payroll locally.

GOOGLE

AlphaGenome Atlas maps DNA changes

The Summary: Google DeepMind has released AlphaGenome Atlas, a map of predictions for all 9 billion possible single-letter DNA changes in the human genome. It predicts what each change may do inside cells and which ones are most likely to matter. Researchers can then focus on the mutations most worth testing in the lab.

Key details:

  • Scientists can use the Atlas to find which DNA mutations may cause disease, before spending months testing them in the lab

  • Each mutation gets a score that helps researchers rank which ones deserve closer study

  • Only about 2% of the genome codes for proteins. The Atlas also reads the other 98%, where many disease-linked changes can hide

Why it matters: Genomics had a search problem: scientists can sequence DNA at scale, but finding the few mutations that matter can take years. AlphaGenome Atlas ranks billions of possible changes and gives researchers a much smaller set to investigate. The next test is whether those rankings keep holding up in experiments. If they do, it will save huge amounts of lab time and accelerate biological research.

TOOLS

🥇 New tools

That’s all for today!

If you liked the newsletter, share it with your friends and colleagues by sending them this link: https://thesummary.ai/