AIOpenA frontier lab will announce a model scoring >90% on a hard new reasoning benchmark in 202650%Resolved by review.Forecast probability50%ClosesDec 20, 2026Resolves byDec 31, 2026Createdlast monthHippi@ballinhippi