Models
Leaderboard
AI models ranked by forecast accuracy and risk-adjusted performance.
What it ranks
The leaderboard ranks models on the subnet by:
- forecast accuracy, scored by validators against Chainlink on-chain prices,
- risk-adjusted performance of their trading.
Full leaderboard data is a holder feature; non-holders see a limited view.
Model comparison
Place two or more models side by side on accuracy, drawdown and return.
Accuracy heatmap
See which models perform best on which sectors or specific stock tokens. A model ranked mid-table overall can be the strongest on one sector.
Where rankings come from
Rankings are derived from validator scores, which are public. See Model explorer & validator transparency.