Tool · Part of the handbook

Polymarket vs Kalshi arbitrage: live price gaps and how to read them

The table below compares live prices for questions listed on both Polymarket and Kalshi. A gap only becomes profit if it survives both venues' fees, the depth at the quoted price, and the chance that the two markets resolve differently. Most visible gaps do not.

Loading live prices…

QuestionKalshi YES bid / askPolymarket YES bid / askMidpoint gap (Polymarket − Kalshi)Cheapest YES + NOEdge after fees
Loading live prices…

Prices are the best bid and ask for one contract that pays $1, read from each venue's public market-data API about once a minute. "Edge after fees" is the payout minus the cost of buying YES on one venue and NO on the other, minus an estimate of each venue's taker fee on a 100-contract order: the market's own Polymarket fee rate, and about 0.07 × contracts × price × (1 − price) for Kalshi, whose exact rounding depends on the account. It ignores depth beyond the best price and any difference in how the two markets resolve. Not financial advice.

How Polymarket and Kalshi arbitrage works

A binary contract pays $1 if its outcome happens and nothing if it does not. If the same question trades on two venues, you can buy YES on one and NO on the other. Exactly one of them pays $1. If the two together cost less than $1, the difference is locked in, before fees.

A worked example with made-up prices:

LegVenuePrice
Buy YESKalshi (ask)$0.42
Buy NOPolymarket (ask, equal to 1 − the YES bid of $0.47)$0.53
Total cost$0.95
Guaranteed payout$1.00
Gross edge$0.05 per contract

Two things to notice. First, the useful prices are the asks you would pay, not the midpoints that most price charts show; that is why the table has a separate "midpoint gap" column, which looks bigger than what you can actually capture. Second, the NO price on one venue is roughly one minus its YES bid, because the two sides of a binary market mirror each other.

The four checks that decide if a gap is real

1. Fees on both legs

Both venues charge takers a fee that is largest near 50 cents and shrinks toward 0 and 100 cents. Two taker fees near the middle of the range can take two to four cents of a five-cent gap. The table subtracts an estimate of each venue's taker fee; the fee guide explains the formulas and the calculator runs your own numbers.

2. Depth at the quoted price

The best ask may be good for a few dozen contracts or for many thousands, depending on the market. Past it you buy at the next price level, and the gap shrinks or reverses. Before trading, read both order books and work out the average price for your full size. The API guide shows how to pull the Polymarket book.

3. Resolution rules

This is where most "free money" disappears. Two markets can ask what sounds like the same question and still resolve differently: a different data source, a different cut-off time, a different treatment of ties, revisions, postponed or cancelled events. If the two markets resolve differently, for example Kalshi settles NO while Polymarket settles YES on the same event, both of your legs lose and you are out the full $0.95. Read both rule pages in full. The table marks pairs whose rules differ in ways we know about; hover or tap the label for the detail.

4. Money, time and eligibility

  • Capital is locked until both markets settle. A two-cent edge on a market that settles in six months is a small return for the time.
  • Funding is separate. Each venue holds its own balance. Moving money between them costs time and sometimes fees.
  • Eligibility differs. Polymarket International restricts users in the US and several other countries to closing existing positions, and Kalshi has its own eligibility rules. Trading both legs is only possible if you may lawfully use both venues. See US vs International.

Arbitrage inside Polymarket

Gaps also appear within a single venue, with no cross-venue risk:

  • Multi-outcome events. In an event where exactly one outcome must win, the YES prices across all outcomes should add to about $1. If buying YES on every outcome costs less than $1 after fees, the set pays $1 whatever happens. If selling every outcome brings in more than $1, the reverse applies.
  • YES plus NO. In a single market, the YES ask plus the NO ask should not fall below $1. When it briefly does, the pair can be bought and redeemed.
  • Related markets. "Above $100k on Friday" cannot be more likely than "above $95k on Friday". When ladders like this are out of order, a spread between them has a bounded loss.

Automated traders watch for these too, so they tend to close quickly. A bot that looks for them needs streaming data rather than polling; the API guide covers the WebSocket feeds.

Building an arbitrage bot

An arbitrage bot is mostly bookkeeping. The order of work that tends to hold up:

  1. Match markets by hand first. Keep a list of pairs you have read the rules for. Automatic matching on titles produces confident mistakes.
  2. Read both books, not the tickers. Compute the average price for your size on each leg, then subtract both fees.
  3. Plan for one leg failing. The second order can miss after the first fills. Decide in advance whether the bot chases, unwinds or holds, and cap how much one-legged exposure it may carry.
  4. Paper trade it. Log every opportunity the bot would have taken and what the book looked like a few seconds later. The risk guide covers paper trading and limits.
  5. Only then automate orders, with small size, loss limits and a kill switch.

Questions people ask

Is Polymarket vs Kalshi arbitrage legal?

Buying on one venue and the opposite outcome on another is ordinary trading, if you are eligible to use both. The usual problem is eligibility: Polymarket International restricts US users to closing positions. Check each venue's terms for your country.

Why do the gaps exist at all?

Different users, different fees, money that cannot move instantly between venues, and real differences in the rules. A large gap that lasts is more likely to reflect one of these than a mistake by the market.

How often does this page update?

About once a minute. Prices are read from each venue's public API and cached for 60 seconds, so a gap you see may already be gone.

Can you add a market?

We add pairs once we have read both venues' rules and they ask the same question. Pairs disappear from the table when either market closes.

Published November 15, 2026. Price sources: Polymarket market data and Kalshi API. Fee sources are listed in the fee guide.

Product waitlist

Follow Bot for Polymarket

We are also developing a no-code Polymarket bot. Leave an email for occasional updates; no wallet connection or trading keys needed.

More guides

We use your email for the updates you request. Privacy policy.