# Sam Altman says an unreleased OpenAI model proved a Millennium Prize Problem

> Original title: The next GPT, which is going to be released soon, was the model OpenAI used to prove a Millennium Prize Problem.

- Source: Rohan Paul
- Published: 2026-11-02T22:07:47.000Z
- Bad Influence: https://badinfluence.news/items/sbqamoe0vw1yf87kugfs5tjs2
- Original: https://x.com/rohanpaul_ai/status/2117378163311562788

## Summary

Rohan Paul relays Sam Altman's remarks at the ADIPEC opening ceremony in Abu Dhabi that the next GPT model, to be released soon, was used to prove a Millennium Prize Problem. The post adds that 10,000 copies of the model ran nonstop for about four days and worked without hierarchy, each choosing its own direction and sharing progress. The full video is on the ADIPEC Official YouTube channel.

## Why it matters

The post quotes Sam Altman's claim about an unreleased model and its reported math result, which is useful context for judging upcoming OpenAI releases, though the claim remains unverified here.

## Post

The next GPT, which is going to be released soon, was the model OpenAI used to prove a Millennium Prize Problem.

~ Sam Altman, at the ADIPEC opening ceremony in Abu Dhabi.

> 10,000 copies of it ran nonstop for about 4 days and proved a Millennium Prize Problem.

> The copies worked with no hierarchy: each chose its own direction, shared progress and learned from the others.

----
Full video on "ADIPEC Official" YouTube channel, (link in comment)

[Video](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Famplify_video_thumb%2F2117377438670082048%2Fimg%2FXVghCm5y-5Djw1Q3.jpg&mode=thumb&exp=1793923200&sig=af338e118d232b04)