How many more Millennium Prize Problems will AI solve in 2026?

The central fact is that the market only resolves Yes if AI solves zero more Millennium Prize Problems in 2026, and the strongest recent development is OpenAI’s public Navier-Stokes claim. That claim is the main reason this market is not trading much higher on Yes, because it is a direct assertion of a full solution to one of the remaining problems, made by an eligible lab, and it could in principle qualify if the public record is strong enough. But as of now it remains unverified, disputed, and not certified by the Clay Mathematics Institute, which matters because the market’s wording and the resolution context both emphasize a qualifying announcement rather than a loose or partial research milestone. On the current evidence, no problem has yet been cleanly and unmistakably solved in the way this market requires.

The other big consideration is timing. There are still a few months left in 2026, and the new OpenAI claim may not be the only catalyst. If one eligible lab has already decided to go public with a bold mathematical breakthrough, that raises the odds of follow-up claims, either from the same lab or from another team trying to establish priority. OpenAI’s mention of substantial progress on another Millennium problem also keeps a second pathway open. Still, converting a claimed result into a qualifying market event requires a public announcement that clearly says the problem is solved, comes from an eligible lab, and is not merely a progress update, a partial result, or a paper that later collapses under scrutiny. That combination is hard to satisfy, and history suggests that major mathematical claims often take time to stabilize, especially when they are extraordinary.

Market pricing is already telling a nuanced story. The current 59% No price implies traders see the no-solution outcome as somewhat more likely, but not overwhelmingly so, which is consistent with a world where one major claim exists but verification risk is high. The fact that the market is not near an extreme suggests participants recognize both sides: the upside case is real because a single accepted announcement would immediately flip the market, while the downside case remains strong because unverified claims often fail to meet the exacting standard needed for a final resolution. My own read is that the biggest edge still sits with zero qualifying solutions, because the path from a claim to a qualifying year-end announcement is narrow and time is limited, but the probability of a qualifying breakthrough is substantial enough that the market should not price Yes too cheaply.