On 13 September 2026, two of our forecasting configurations assigned about 98% probability to a YES resolution on Manifold’s question about whether Navier-Stokes would be considered solved by 3 June 2027. Four days later, the market resolved YES.

The median forecast across the 113 eligible entries in the same ForecastBench round was approximately 32%. Our entries, listed as Frost and Linen, both recorded 97.82%.

Forecasts submitted 13 September; market resolved YES on 17 September
Forecast / comparisonProbability of YES
Ryso: Frost97.82%
Ryso: Linen97.82%
Ryso: Apex90.86%
Median of 113 eligible entries31.64%

Source: ForecastBench public forecasts, with the median calculated by RysoAI. CC BY-SA 4.0. CC BY-SA 4.0

The question being forecast

The market asked whether mathematicians would broadly consider the problem solved by its deadline. ForecastBench used the market’s resolution as the outcome to predict. On 17 September, the creator resolved it YES, stating that mathematicians widely agreed the proof was valid.

OpenAI announced its result on 8 September. Five days later, our models had to judge whether the proof would win broad acceptance by the market’s deadline.

Explore RSO-F

From the archive. The date refers to the period described.

Original Manifold question and YES resolution

ForecastBench September 13 question set

ForecastBench September 13 resolutions

Public processed forecasts

OpenAI’s September 8 announcement

Our October 12 case study