Prediction market leaderboard: open interest, volume, turnover.

Questions about this leaderboard

Which prediction market has the most open interest right now?

The leaderboard below ranks every tracked venue by open interest, read from each venue's public API or from the DefiLlama aggregate.

What does turnover mean on this page?

24-hour volume divided by open interest: how many times a venue's book turns over in a day. A low figure means capital parked on long-dated outcomes, a high one means short-dated markets on a thin book.

Where do the open interest and volume numbers come from?

Open interest comes from the venue where the venue publishes a usable figure and from DefiLlama protocol TVL otherwise, and for open interest that is more often DefiLlama: Kalshi publishes it natively, while Polymarket's gamma field is deprecated and reads an order of magnitude low, so Polymarket and Limitless use TVL. Volume is each venue's own where its API gives one, otherwise DefiLlama. Where a venue's own read is a partial view the aggregate wins: unauthenticated, the harness can only sum Kalshi's event catalog and reaches $2.26M against DefiLlama's $424.8M for the same day, so Kalshi's volume is the aggregate's. Every row's tag names its own source.

Why do some venues show no latency or resolution figure?

Those columns come from separate probes (pm-api-latency, pm-resolution-delay, pm-ws-latency) which run against venues with a documented public API. A venue fed through the DefiLlama aggregate carries open interest and volume but no probe columns, and shows a dash rather than a zero.

Can I cite these numbers?

Yes. Every figure is reproducible from public sources, released under CC BY 4.0, and the bench pages expose machine-readable /api/stat endpoints with the same values and timestamp.