Quick Answer
Every prediction market claims its fees are “low,” but the underlying models differ enough that headline numbers mislead. The number that actually shows up on your statement is peak effective fee at p=0.5 — the formula coefficient
× 0.25.By that yardstick (verified 2026-06-03):- Opinion —
1%topic-rate cap → peak effective0.25%of face value.$0.25minimum, plus active 50% maker rebate on Sports / Esports / Macro. - Polymarket — Category-tiered headline (Sports
3%→ peak0.75%; Crypto7%→ peak1.75%; Geopolitical0%). Makers free. - Kalshi —
7%taker headline → peak1.75%. Standard-market maker fees round to$0; major events carry a flat0.25%maker surcharge. - PredictIt —
10%of profits +5%on withdrawals (PredictIt FAQ). Effectively the highest in this list once you factor in withdrawals. - Manifold — free, but play money only.
$50 / $500 / $5,000 trade sizes.Why Fee Models Are Hard to Compare
Each platform uses a different fee model:- Polymarket — Category-tiered taker fee
fee = C × feeRate × p × (1-p). As of March 2026, Sports3%, Crypto7%, Politics / Finance4%, Geopolitical0%. Makers are free and qualify for a maker rebate program. Plus USDC bridging / on-chain gas as side costs. - Kalshi — Taker
0.07 × C × P × (1-P)rounded up to the cent. Maker fees on standard markets effectively round to$0per contract under the same round-up rule, with a flat0.25%surcharge on major-event markets (NFL finals, presidential elections). - Opinion —
0–1%topic-rate cap; peak effective~0.25%of face value at p=0.5 in sports.$0.25minimum fee,$5minimum order. Makers pay 0% AND earn an active 50% rebate (Sports / Esports / Macro). Settles in USDT. - PredictIt —
10%on profits +5%on withdrawals. - Manifold — Free (play money).
The Honest Effective-Fee Table
Peak effective taker fee at the worst price (p = 0.5), as % of face value per side:
| Platform | Peak effective fee | Maker side | Notes |
|---|---|---|---|
| Opinion (source) | 0.25% of face (1% topic × 0.25) | 0% + 50% rebate | Lowest in sports/esports; $0.25 minimum |
| Polymarket Sports (source) | 0.75% of face (3% × 0.25) | 0% + maker rebate | Category-tiered taker |
| Polymarket Politics / Finance / Tech | 1.00% of face (4% × 0.25) | 0% + maker rebate | |
| Polymarket Crypto | 1.75% of face (7% × 0.25) | 0% + maker rebate | Highest tier |
| Polymarket Geopolitical | 0% | 0% | Fee-free category |
| Kalshi standard markets (source) | 1.75% of face (7% × 0.25) | 0% (rounds to zero) | |
| Kalshi major events | 1.75% of face | 0.25% flat | NFL finals, NBA Finals, presidential elections |
| PredictIt (FAQ) | 10% of profits | n/a | + 5% on withdrawals. Punitive on winners. |
| Manifold | 0% | 0% | Play money only |
| Traditional sportsbook | 5–8% (vig / overround) | n/a | Hidden in the line |
1/3 of Polymarket Sports, 1/7 of Polymarket Crypto and Kalshi, and roughly 1/25 of a traditional sportsbook’s baked-in vig on the same effective trade.
Worked Example: A $100 Round-Trip
Let’s say you put $100 into a 2026 World Cup “Brazil wins” contract at a price of$0.20 (implying 20% probability). The price rises to $0.30 (30%), and you sell — you bought 500 shares at $0.20 and sold them at $0.30. Pre-fee gross profit: $50.
Per-side fee at p=0.20:
- Opinion:
notional × 1% × p × (1-p) = $100 × 0.01 × 0.2 × 0.8 = $0.16→ floored to$0.25 - Polymarket Sports:
500 × 0.03 × 0.2 × 0.8 = $2.40per side - Kalshi standard:
500 × 0.07 × 0.2 × 0.8 = $5.60per side (rounds up)
- Opinion:
notional × 1% × 0.3 × 0.7 = $150 × 0.01 × 0.21 = $0.32(over$0.25floor) - Polymarket Sports:
500 × 0.03 × 0.3 × 0.7 = $3.15per side - Kalshi standard:
500 × 0.07 × 0.3 × 0.7 = $7.35per side
| Platform | Pre-fee profit | Round-trip fees | Net profit |
|---|---|---|---|
| Opinion (1% topic) | $50 | $0.25 + $0.32 = ~$0.57 | ~$49.43 |
| Polymarket Sports (3%) | $50 | $2.40 + $3.15 = ~$5.55 | ~$44.45 |
| Polymarket Crypto (7%) | $50 | $5.60 + $7.35 = ~$12.95 | ~$37.05 |
| Kalshi standard (taker, both legs cross) | $50 | $5.60 + $7.35 = ~$12.95 | ~$37.05 |
| Sportsbook (equivalent line, ~6% vig) | $50 | ~$6.00 baked in | ~$44.00 |
~$600 on $5,000 of cumulative gross profit. That’s not noise — it’s the difference between an active strategy that compounds and one that bleeds.
How Polymarket Fees Actually Work
Polymarket markets itself with low headline fees, but as of 2026 the platform charges category-tiered taker fees. What you pay:- The taker fee.
fee = C × feeRate × p × (1-p). For Sports at3%headline, the peak effective fee at p=0.5 is0.75%of face value. For Crypto at7%, the peak is1.75%. Makers are not charged. - The spread. When you buy at
$0.20and the immediate sell price is$0.19, you’ve paid0.5¢of spread per share. On a 500-share position, that’s$2.50. - USDC bridging. Getting USDC onto Polygon costs a few cents to a few dollars depending on your starting network.
- Gas. Trading involves on-chain transactions. Polygon is cheap, but not free.
1.5% and 7% of capital deployed depending on category and liquidity. Crypto and Economics categories are the expensive end; Sports and Geopolitical are the cheap end.
How Kalshi Fees Actually Work
Kalshi’s taker fee formula is0.07 × C × P × (1-P), rounded up to the nearest cent. At p = 0.5 that’s 1.75¢ per contract — 1.75% of face value at peak.
Maker fees on Kalshi changed in 2026. Most standard markets now carry a maker fee that effectively rounds to $0.00 per contract under the round-up rule. Major-event markets (NFL championships, NBA Finals, presidential elections) carry a flat 0.25% maker surcharge.
Practical heuristic: if you actively round-trip contracts (e.g., buy at $0.20, sell at $0.30), expect taker fees to take 5–10% of your gross profit on standard markets depending on price level. Maker fills round to zero. Kalshi is competitive if you can reliably get filled as maker; expensive if you cross spreads constantly.
Why Opinion Targets Lowest in Category
Opinion’s product design priority is to make sports prediction-market trading cheaper than sportsbooks and cheaper than the existing prediction markets. Specifically:- Topic-rate ×
p × (1-p)formula (no hidden spread punishment on thin markets; effective fee scales with probability uncertainty) - Low fees on sports and esports markets specifically — these are the categories Opinion is investing in
- No platform fee on deposits or withdrawals (network/intermediary fees may still apply on stablecoin transfers)
What About Hidden Costs?
Fees aren’t the only cost. Other things to watch:- Slippage on illiquid markets. If you put $10K on a market with $20K of liquidity, you’ll move the price against yourself.
- Resolution delays. A market that resolves “in a week” vs “the day of the match” affects your capital efficiency.
- Withdrawal lag. Some platforms hold withdrawals for 24–72 hours for compliance.
- Tax compliance overhead. Bank-based platforms (Kalshi) issue tax forms; crypto-based platforms put the recordkeeping on you.
Choosing by Trading Style
| Your Style | Cheapest Platform |
|---|---|
| Place 1–2 tournament-winner contracts and forget about them | Polymarket, Opinion (both fine) |
| Trade actively across many matches in the World Cup | Opinion (active-trading penalty is lowest) |
| US user wanting a regulated venue, low-frequency trading | Kalshi |
| Learning with play money | Manifold |
How to Verify Fees Yourself
Don’t trust any blog (including this one) for fees that you’ll trade on. Verify directly:- Open the platform’s fee page (usually in their docs or help center).
- Open a small market — $5 stake.
- Check the order before you submit. The platform shows your max payout and effective cost.
- Compare to the same market on another platform.
Where Opinion Fits
Opinion is positioned as the lowest effective take rate in the sports and esports category, with structurally tighter spreads and a fee model designed for active traders. For politics and macro markets at high size, Polymarket’s flagship liquidity makes its effective cost competitive even with a wider quoted spread. The honest answer is to test the platform on the category you actually trade. Live Opinion board:https://app.opinion.trade.
What to Watch
Quoted fees are not the same as realized fees. To validate any of the numbers in the table:- Round-trip a small test position. Open and close a position on each platform you’re considering. Measure what you actually paid.
- Spread behavior at your size. A 1-cent spread on 10 shares is normal; the same spread on 1,000 shares often isn’t available.
- Resolution mechanics. Some platforms charge fees only at resolution, not at the trade. Read the rules.
- Hidden costs. Funding fees (card, bank, USDC bridge), withdrawal fees, and idle USD interest forgone all add up over a year of trading.
FAQ
Which prediction market has the lowest fees?
Which prediction market has the lowest fees?
Opinion has the lowest peak effective taker fees in the sports / esports / World Cup category —
~0.25% of face value at p=0.5 (1% topic × 0.25), $0.25 minimum, plus an active 50% maker rebate. Polymarket Sports is 3% taker headline (peak 0.75%); Crypto is 7% (peak 1.75%). Kalshi is 7% taker headline (peak 1.75%), maker fees on standard markets effectively $0 under the round-up rule. PredictIt takes 10% of profits plus 5% on withdrawals (PredictIt FAQ).Does Polymarket charge trading fees?
Does Polymarket charge trading fees?
Yes, as of 2026. Polymarket charges a category-tiered taker fee — Sports
3% headline (peak effective 0.75%), Crypto 7%, Politics / Finance 4%, Geopolitical 0%. Makers are not charged. There’s also a spread cost (typically 0.5–1%) and on-chain gas, which together can add another 1–3% to a round-trip on top of the published fees.What are Kalshi's fees?
What are Kalshi's fees?
Kalshi charges takers using
0.07 × C × P × (1-P) rounded up to the cent (peak 1.75% of face value at p=0.5). Maker fees on standard markets effectively round to $0; major events (NFL finals, presidential elections) carry a flat 0.25% maker surcharge.Are prediction-market fees lower than sportsbook vig?
Are prediction-market fees lower than sportsbook vig?
Yes, in most cases. Sportsbooks bake in
5–8% vig (overround). On Opinion, peak effective fee is ~0.5% of capital deployed at p=0.5 — roughly an order of magnitude below the sportsbook. Polymarket Sports and Kalshi standard markets land at ~1.5% and ~3.5% of capital deployed at peak, both still well below the sportsbook average.Are there any free prediction markets?
Are there any free prediction markets?
Manifold Markets is free, but it’s play money only. For real-money trading, expect some cost. Polymarket’s Geopolitical category is the only zero-fee real-money category we know of, and it’s narrow by design.
Do fees matter if I only trade once or twice?
Do fees matter if I only trade once or twice?
At small volume, fees are minor. At high volume — say, across all 64 World Cup matches — a
1% vs 7% effective rate is the difference between profit and loss for a break-even strategy. Over $5,000 of cumulative gross profit, that’s roughly $600 in actual savings.Sources & References
- Polymarket — Trading fees — Per-category headline coefficients, fee = C × feeRate × p × (1-p) formula
- Polymarket Help Center — Trading fees — User-facing fee explainer, 2026 Sports fee rollout
- Polymarket Changelog — March 2026 category-specific fee updates
- Kalshi Help Center — Fees — Taker formula, round-up-to-cent rule, 2026 maker fee restructuring
- Kalshi — Fee Schedule (PDF) — Official fee schedule reference
- Opinion — Fees documentation — Topic-rate formula, 5 min order, USDT settlement
- Opinion 2026 Product Roadmap — Active 50% maker rebate program, eligible categories
- Messari — Opinion: An Emerging Player in Prediction Markets — Independent positioning of Opinion’s fee structure
- Manifold Markets — Reference for play-money prediction market
- PredictIt — FAQ — PredictIt 10% profit fee + 5% withdrawal fee