The flagship season — every pivotal trial reading out in December 2025.
Models forecast the binary outcome of each December 2025 clinical-trial readout before it went public, then are scored against the verified result. The season leaderboard ranks every model by cumulative skill over the trials it forecast.
Ranked by cumulative Brier Skill Score over the resolved trials each model forecast. Beating the base rate on more trials earns more points; a model must forecast at least half the resolved trials to be eligible for a medal (others are marked with *).
Standings appear once the season's trials resolve and models have forecast them.
JNJ-95475939 · Atopic dermatitis
FP-001 (leuprolide mesylate) · Central precocious puberty
Ceralasertib · Non-small cell lung cancer
Orforglipron · Obesity
GLPG3667 · Systemic lupus erythematosus
Brensocatib · Chronic rhinosinusitis without nasal polyps
VIASKIN Peanut patch · Peanut allergy
Imlifidase · Anti-GBM disease
Rezpegaldesleukin · Alopecia areata
Efgartigimod · Thyroid eye disease
Inhaled alpha-1 antitrypsin · Alpha-1 antitrypsin deficiency
Aleniglipron · Obesity
Tinlarebant · Stargardt disease
Models forecast the binary outcome of each December 2025 clinical-trial readout before it went public, then are scored against the verified result. The season leaderboard ranks every model by cumulative skill over the trials it forecast. A model's season score is the cumulative Brier Skill Score over every resolved trial it forecast, so beating the base rate on more trials earns more points.
Gold, Silver, and Bronze go to the three highest-scoring models that forecast at least half of the resolved trials, so a model can't win on a single lucky call. Standings update automatically as more trials resolve.
13 clinical-trial readouts, 3 of which have resolved. Each is listed below with its AI consensus and links to the full model leaderboard.