Every prediction market platform wants deep order books, and none of them get liquidity for free. The four platforms below pay for it in different ways — cash rebates, points, reduced fees, or nothing at all for retail users. Here's what each one actually offers as of mid-2026, and how the payouts hold up when you split them by market direction.
A note before the numbers: every platform reserves the right to change these terms without much warning. Treat this as a snapshot, not a contract.
Kalshi is the outlier here. It runs institutional rebate programs — a block-trade rebate and a sportsbook-hedging rebate, both filed with the CFTC — but neither one is something a retail trader can sign up for. For everyday users, the closest thing Kalshi offers is the Liquidity Incentive Program, which pays for keeping size at the best bid or ask, not for fills.
predict.fun sits at the other extreme: no cash changes hands for providing liquidity. Market-making activity (limit orders that get filled) earns Predict Points at a higher rate than simply taking price, and liquidity provided on markets priced away from consensus (what they call "high-uncertainty") earns a 2x points multiplier. Whether those points ever convert to something spendable hasn't been confirmed by the platform.
Up markets vs. Down markets
Several of these venues run short-window "Up or Down" style crypto contracts — Polymarket's 15-minute and hourly crypto markets, Limitless's Hourly and 15-minute Crypto categories, Kalshi's hourly crypto contracts, and predict.fun's crypto markets. Since these pair an "Up" contract against a "Down" contract on the same underlying move, it's worth asking whether the rebate math treats both sides the same.
Polymarket. The fee (and therefore the rebate funded by it) is calculated off a formula that's symmetric around a 50-cent price. A maker resting an order at 30 cents on the Up side generates the identical rebate as a maker resting at 70 cents on the Down side of the same market, because the two prices sum to a dollar and the formula treats them as mirror images. Direction doesn't matter here — distance from the midpoint does.
Limitless. Same underlying logic. Rebate credit comes from a share of the taker fee generated by a fill, and the fee schedule doesn't distinguish Up from Down — it's priced off the contract's own price, and Up/Down prices are complementary. A market maker quoting both sides of an Hourly Crypto market earns rebates on whichever side gets hit, with no direction getting preferential treatment.
Kalshi. There's nothing to compare on the retail side, since there's no public rebate formula to inspect. Whatever asymmetries might exist inside a private Market Maker Agreement aren't disclosed.
predict.fun. Points scale with dollar volume traded as a maker, not with which side of the contract you're on. The 2x multiplier is tied to how uncertain the market is (bid or ask sitting far from consensus), not to whether that liquidity sits on the Up or the Down side. So functionally, an Up-side limit order and a Down-side limit order of equal size at equivalent distance from the midpoint earn the same multiplier.
The upshot: on every platform that publishes its formula, Up and Down markets are rebated symmetrically. The rebate cares about how close to the coin-flip price your liquidity sits, not which direction you're betting will win.
Bottom line
If cash rebates are the goal, Limitless currently pays out the largest share of taker fees on its short-window crypto markets, and Polymarket offers the most mature, best-documented program with two parallel incentive tracks. Kalshi's rebates exist but aren't open to retail traders. predict.fun substitutes points for cash, which may or may not turn into real money later — worth keeping in mind if you're optimizing for near-term yield rather than a speculative future payout.
The programs, side by side
| Feature | Polymarket | Kalshi | Limitless | predict.fun |
|---|---|---|---|---|
| Core rebate mechanism | Maker Rebates: a cut of the taker fee, paid to the resting order that got filled | No public cash-rebate program for retail; discounted fees available only through signed Market Maker Agreements | Maker Rebates: pro-rata share of the taker fee, paid to whoever's resting order got hit | No cash rebate. Predict Points for market-making activity instead |
| Rebate size | 15–25% of the taker fee, depending on category (crypto and sports currently run lower than the rest) | Not published — negotiated per market maker | Currently 100% of the eligible taker fee on Daily, Hourly Crypto, and 15-minute Crypto markets | Points-based, not a fee share — no fixed percentage |
| Payout currency/form | pUSD, credited automatically | Fee waiver / reduced schedule, not a cash payout | USDC | Predict Points (possible future token, unconfirmed) |
| Payout timing | Daily, $1 minimum before payout triggers | N/A for retail | Rebate credit accrues continuously, USDC paid out once per day (UTC) | Weekly, with a 2–3 day calculation lag |
| Separate "just for resting" program | Yes — Liquidity Rewards, pays for orders sitting near the midpoint whether or not they fill | Liquidity Incentive Program (Sept 2025–Sept 2026): pays for maintaining best bid/ask presence, fill or no fill | Yes — LP Rewards, pays for proximity to midpoint and order size, fill or no fill | No separate program; points already cover both filled and resting activity loosely |
| Who can access it | Any wallet, no application | Institutional/professional entities only, via application and agreement | Any wallet, no application | Any account, no application |
| Fees makers pay to trade | Zero | Sometimes a maker fee applies, market-dependent | Zero | Zero (taker-only fee model) |
Official program pages
- Polymarket — Maker Rebates, Liquidity Rewards
- Kalshi — Fee Schedule, Liquidity Incentive Program, Becoming a Market Maker
- Limitless — Maker Rebates, LP Rewards
- predict.fun — How to Earn Points
Rates, tiers, and eligibility on all four move around often enough that it's worth checking the source before sizing a strategy around any single number in this article.