AI Milestones 2026—AGI, Millennium Prize and releases

The two Astra (6.1) release books now agree on the year. Kalshi prices before December 31 at 98.5% and Polymarket by December 31 at 97.95%; mid-November is closer to a coin flip, with Kalshi's before November 13 at 48.5% and Polymarket's by November 13 at 55% and by November 15 55.5%, on $26,077 and $21,714 of daily volume.

The Manifold anchor stays low. The Navier-Stokes contract—will Anthropic announce a solution before its IPO—sits at 3.86%, with a November 30 close.

The multi-year boards frame it. Kalshi's Millennium Prize board: before 2028 4%, before 2029 9.75%, before 2030 58.95%, before 2035 83.75%, with nothing traded in the last day. The AGI pair stays low too: Kalshi's OpenAI achieves AGI before 2027 at 4% against Polymarket's announces before 2027 at 6%.

The Manifold anchor and its broader siblings. The Navier-Stokes contract asks a narrower question than the others: not whether an AI solves the problem, but whether Anthropic announces it, before an IPO. It is 3.86%, and it closes on November 30. The sibling—AI solves another Millennium problem, besides Navier-Stokes, before September 15, 2027—prices 83.06%, and Manifold's AI solves a Millennium Prize Problem before 2030 is 95.32%.

Kalshi's ladder puts the next Millennium Prize being awarded before 2027 at 1.5% and before 2030 at 58.95%, while Manifold prices an AI solving another problem by September 2027 at 83.06%. Solving and awarding are different events separated by the Clay Institute's two-year publication requirement, which is most of the gap—but not obviously all of it.

Polymarket asks whether OpenAI announces AGI before 2027 (6%, on $36 in the last day); Kalshi asks whether OpenAI achieves it (4%). Both resolve on public statements in practice, and the two sit within two points. Kalshi's ladder steps to 53.6% before 2028 and 68.65% before 2030, on $402k of open interest.

Left: Kalshi's Astra ladder: before November 13 at 48.5% and before December 31 at 98.5%, on $26,077, $21,348 of it on the year-end rung. Right: Polymarket's Astra (6.1+) board: November 13 55%, November 15 55.5%, November 30 93.25%, December 15 97% and December 31 97.95%, on $21,714, $4,998 of it on the November 15 leg. For the end of the year the two venues sit within a point; for mid-November, Polymarket's November 13 leg is 6.5 points above Kalshi's.

Polymarket's Luna (6.1+) ladder—November 15 at 8%, November 30 at 58.5%, where all $402 of the day's trade went, and December 31 at 80%—beside Kalshi's any-company AGI board, which turned $14,189. Its long rungs are nearly flat: before July 2028 at 60%, before October 2028 62%, before 2029 76%, before 2030 79% and before 2031 77.5%. The 2031 rung still sits below the 2030 one, so read it as a stale quote.

The Navier-Stokes contract is at 3.86%, against Manifold's own 83.06% that AI solves a different Millennium problem by September 2027. The market doubts the specific announcement, not the capability.

Astra 6.1's near dates are the open question. The year-end rungs are 98.5% on Kalshi and 97.95% on Polymarket, but Kalshi's before November 13 is 48.5%, and Polymarket's November 13 and 15 legs are 55% and 55.5%.

The AGI pair stays low—4% on Kalshi and 6% on Polymarket for 2027, with both contracts running to the end of the year.

Luna by the end of November, at 58.5%, took all of that board's trade in the last day, while its November 15 leg is 8%.