Which Millennium Prize Problem will AI solve next?
This event asks which Millennium Prize Problem an eligible AI lab will first announce as solved by December 31, 2027, 11:59 PM ET. Five problems qualify; Navier-Stokes does not. The key rule is that the lab itself must announce the solution as its own work, and no confirmation from the Clay Math Institute or anyone else is needed.
Leading right now: Riemann Hypothesis 82%, Hodge Conjecture 9%, Yang-Mills 6%, No solution by Dec 31, 2027 3%.
| Market | Chance of yes | Deadline |
|---|---|---|
| Will the Riemann Hypothesis be the next Millennium Prize Problem solved by AI? | 81.8% | Jan 1, 2028, 4:59 UTC |
| Will the Hodge Conjecture be the next Millennium Prize Problem solved by AI? | 8.7% | Jan 1, 2028, 4:59 UTC |
| Will the Yang-Mills existence and mass gap problem be the next Millennium Prize Problem solved by AI? | 6.2% | Jan 1, 2028, 4:59 UTC |
| Will no Millennium Prize Problem be solved by AI by December 31, 2027? | 2.9% | Jan 1, 2028, 4:59 UTC |
| Will the Birch and Swinnerton-Dyer Conjecture be the next Millennium Prize Problem solved by AI? | 1.4% | Jan 1, 2028, 4:59 UTC |
| Will P versus NP be the next Millennium Prize Problem solved by AI? | 1.1% | Jan 1, 2028, 4:59 UTC |
Odds read from Polymarket at 22:18 UTC on Nov 21, 2026. $525K traded so far, $109K in the last 24 hours.
How this market resolves
Yes: For a given problem's market, YES requires that problem to be the first qualifying one an eligible AI lab announces as solved by the deadline. The announcement must come from the lab or its official representative, say the problem is solved, and present the work as the lab's own, alone or jointly with outside researchers. If two or more problems are announced on the same ET calendar day, the one first in alphabetical order wins. For the 'no problem solved' market, YES requires that no such announcement is made by the deadline.
No: A problem's market is NO if a different qualifying problem is announced first, or if no qualifying announcement is made by the deadline. The 'no problem solved' market is NO if any qualifying announcement is made by the deadline. A same-day tie goes to the alphabetically first problem, so the others resolve NO.
When it really ends
December 31, 2027, 11:59 PM ET (Eastern Time, UTC-5 in winter), which is January 1, 2028, 04:59 UTC. The endDate field in the data (2028-01-01T04:59:00Z) matches the rules.
What counts
- An announcement by an eligible lab or its official representative that one of the five listed problems has been solved.
- Work presented as done by the lab, its researchers or its models, alone or jointly with outside researchers.
- Announcements from subsidiaries and research divisions of listed labs.
- Announcements from any other company with a model on the MMLU-pro leaderboard at the time of the announcement.
- A consensus of credible reporting, which the rules say may also be used as a resolution source.
What does not count
- Any announcement about the Navier-Stokes existence and smoothness problem.
- Partial results or progress toward a solution.
- Announcements that only credit or congratulate outside researchers, even if they used the lab's models, compute or research credits.
- Announcements by companies not on the list and not on the MMLU-pro leaderboard, or by other organizations or individuals.
- Announcements made after December 31, 2027, 11:59 PM ET.
Fine print people miss
- No confirmation from the Clay Math Institute or any other body is required, so an announcement alone can be enough, whether or not the proof is verified.
- Ties on the same ET calendar day go to the problem first alphabetically: Birch and Swinnerton-Dyer, Hodge, P versus NP, Riemann, Yang-Mills.
- The rules do not say what happens if an announcement is later retracted or found to be wrong.
- Eligibility is judged at the time of the announcement, and the leaderboard-based clause makes the lab list open-ended.
- The rules do not say how the 'first' announcement is decided within a single day; only the alphabetical tie-break applies across problems.
What moves the price
- A lab announcing a claimed solution to a listed problem would move that problem's market up and the 'no solution' market down.
- A lab publishing partial results or progress would not count, but could change how likely people think a later full claim is.
- Credible retractions or errors found in a claimed proof matter only if the rules treat the announcement as still standing, which they do not address.
- Announcements about Navier-Stokes would not move any market directly, since that problem does not qualify.
- Announcements crediting outside mathematicians rather than the lab would not count and so should not push any market toward YES.
Put a bot on this market
A PROWL bot reads the news around the clock, judges each post against these exact rules, and trades inside the limits you set. It starts in paper mode.
Questions people ask
Which Millennium Prize Problems count in this market?
Five count: the Birch and Swinnerton-Dyer Conjecture, the Hodge Conjecture, P versus NP, the Riemann Hypothesis, and the Yang-Mills existence and mass gap problem. Navier-Stokes is excluded.
Does the Clay Math Institute have to confirm the solution?
No. The rules say no further confirmation from the Clay Math Institute or any other organization is required. An eligible lab's announcement is enough.
What if an AI helps a human mathematician solve a problem?
It counts only if the lab's announcement presents the solution as the work of the lab, its researchers or its models, alone or jointly with outsiders. A statement that only credits outside researchers does not count.
When does the market end?
On December 31, 2027 at 11:59 PM Eastern Time, which is January 1, 2028 at 04:59 UTC.
What happens if two problems are announced on the same day?
If both are announced on the same ET calendar day, the market resolves to the one that comes first alphabetically in the rules' list.
Which companies are eligible AI labs?
The rules list many labs, including Anthropic, Google, OpenAI, Meta, DeepSeek and Microsoft. Any other company with a model on the MMLU-pro leaderboard at the time of the announcement also qualifies.
More Science markets explained
Written from the market's own rules text on Nov 16, 2026. If the rules on Polymarket differ from this page, the rules on Polymarket govern.