← All posts · June 24, 2026 · 7 min read

Polymarket taker fees, explained with real dollar math

Polymarket ran fee-free for years. That ended in January 2026, quietly, and if you trade the fast crypto markets — or copy anyone who does — the fee formula is now part of your P&L whether you've read it or not.

Most traders haven't. Let's fix that with actual numbers.

How fees arrived: a short timeline

Polymarket didn't flip one big switch. It rolled fees out market type by market type, and you can trace the whole thing through its public changelog:

• January 5–6, 2026: taker fees go live on 15-minute crypto up/down markets. Cointelegraph called it a quiet introduction — no splashy announcement, just updated docs.
• February 12, 2026: the 5-minute crypto markets launch, with taker fees enabled from day one.
• March 6, 2026: fees extend to all crypto markets — hourly, 4-hour, daily, weekly.
• March 30, 2026: "Fee Structure V2" brings category-based rates to sports, politics, finance and the rest. Geopolitics stays fee-free.

Why fees at all? Not revenue, at least not directly. As The Block reported, collected taker fees get redistributed daily to market makers as liquidity rebates. The fast crypto markets were getting picked apart by latency arbitrage — bots hitting stale quotes the moment Bitcoin moved — and charging takers while paying makers is the standard medicine for that.

The formula

Per Polymarket's fee docs, a taker on a crypto market pays:

fee = shares × 0.07 × price × (1 − price)

Three things fall out of that little expression.

First, only takers pay. If your order rests in the book and someone fills it, you pay nothing — makers are never charged, and on crypto markets they actually receive 20% of collected taker fees back as rebates.

Second, the fee peaks at 50/50 and dies at the extremes. The price × (1 − price) term is largest at $0.50 and shrinks toward zero at $0.01 or $0.99. A coin flip is the most expensive thing you can buy; a near-certainty is almost free.

Third, it's per share, not per dollar — which makes cheap shares proportionally pricier than they look. Worked examples, 100 shares each:

• At $0.50: 100 × 0.07 × 0.50 × 0.50 = $1.75 on a $50 position — 3.5% of what you spent.
• At $0.30: 100 × 0.07 × 0.30 × 0.70 = $1.47 on a $30 position — about 4.9% of outlay.
• At $0.90: 100 × 0.07 × 0.90 × 0.10 = $0.63 on a $90 position — 0.7%.
• At $0.99: 100 × 0.07 × 0.99 × 0.01 ≈ $0.07. Basically nothing.

One footnote for precision nerds: at the January launch the curve peaked at 1.56 cents per share (per the changelog), which corresponds to a slightly lower rate than today's documented 0.07. The rate has been tuned since. Always check the docs page for the current number; the shape of the curve is what stays constant.

The trade this killed: both sides at 99 cents

Before fees, there was a beloved near-free-money pattern in the fast markets: buy YES and NO in the same market whenever the two prices summed to $0.99 or less. A matched pair always redeems for exactly $1.00, so a $0.99 pair locked in a guaranteed penny — scale it up, repeat every few minutes, and it added up.

Run that trade through the formula now. Say you take YES at $0.50 and NO at $0.49. Fee on the first leg: $0.0175 a share. Fee on the second: 0.07 × 0.49 × 0.51 ≈ $0.0175. Total cost about 3.5 cents per pair to capture a 1-cent payoff. The trade didn't get worse — it got mathematically dead, exactly as intended, because both legs sit at the top of the fee curve.

It survives in one form: as a maker. Rest limit orders on both sides and get filled passively, and you pay nothing. But then you're doing real market making, with real inventory risk, competing with professionals for fills. The free penny is gone.

What fees change if you trade the 5-minute markets

A lot, honestly — more than most participants price in. People wave fees off as small. At $0.50 in a 5-minute market they are your single biggest fixed cost, bigger than the spread on a liquid book.

Buy the $0.50 side and you need to be right about 51.75% of the time just to break even on the fee alone — before spread, before slippage. Buy at $0.60 and win: you make $0.40 a share minus about 1.7 cents of fee, roughly 4% of your profit gone. Do that 40 times a day and fees quietly become your biggest counterparty. It's also one more reason to be picky about entries instead of hammering the mid every round — patient entries at better prices pay less fee and need less accuracy. The mechanics of these markets have their own quirks; the 5-minute markets explainer covers them properly.

What fees change if you copy trade

Copies are taker orders almost by definition — the whole point is matching someone's entry now, not waiting in the book. So on crypto markets, every copy you fire pays the formula, layered on top of the slippage between their price and yours (we broke that down in the slippage post).

That stacks the math against copying high-frequency crypto wallets in particular: lots of trades near the middle of the curve, small per-trade edges, full taker fee on each one. A wallet that wins 2 cents a trade on its own fills can be a loser in your account on fees alone. It's a big piece of why the profitability question comes down to copying a few big slow edges rather than many small fast ones.

Worth saying because people conflate them: these are Polymarket's fees, not your bot's. Percentage-fee copy bots charge on top of them, per trade. btc5min charges a flat $15 a month and adds nothing per trade, which matters more the more crypto-market copies you fire — but nothing any tool does makes the exchange-side taker fee go away. Only being a maker does that.

btc5min is an independent tool and is not affiliated with, endorsed by, or operated by Polymarket. Nothing here is financial advice. Prediction markets are risky, copying another trader does not guarantee profit, and you can lose the money you trade with.