Kalshi vs Polymarket prices: we compared 351 identical outcomes to see where each exchange is cheaper
We matched 351 outcomes listed on both Kalshi and Polymarket across 31 events and compared their prices at the same moment. In liquid markets the median gap was 1 point, Kalshi was cheaper for “Yes” slightly more often, and Republicans were priced higher on Kalshi in 11 of 15 political races. The data, the crossed books and what it means for traders.
In this article
For the same outcome, Kalshi and Polymarket usually agree to within a point. We matched 351 outcomes listed on both exchanges, across 31 events in politics, economics and sports, and took both order books at the same moment on 30 October 2026. Among the 91 outcomes with a tight two-sided market on both sides, the median price gap was 1.0 percentage point, and only one gap was larger than 5 points.
But small gaps are not the same as no gaps. In 39 of those 91 outcomes, the two order books were "crossed": you could buy "Yes" on one exchange and "No" on the other for less than $1 in total, before fees. And in politics, one pattern stood out: Republican candidates were priced higher on Kalshi than on Polymarket in 11 of 15 races.
This study explains how we did it, what we found and what it means if you trade on both.
Key findings
Source: Our analysis of the Kalshi and Polymarket public APIs, 30 October 2026, 15:48 UTC
- The median gap between the two exchanges' mid-prices was 1.0 point; the average was 1.37 points.
- 42.9% of liquid outcomes were within 1 point; 28.6% differed by more than 2 points; 1 differed by more than 5.
- "Yes" was cheaper on Kalshi in 36 outcomes and on Polymarket in 31; 24 were within half a point.
- Both exchanges had a median bid-ask spread of 1¢; Kalshi's book was tighter in 33 outcomes, Polymarket's in 22, and they were level in 36.
- Gaps were smallest in the busiest markets and largest in the least active third.
- Republican outcomes were 1.5 points higher on Kalshi on average across 15 political races.
How we did it
| Snapshot | 30 October 2026, about 15:48 UTC |
| Exchanges | Kalshi, and Polymarket's international site (polymarket.com) |
| Events | 33 pairs listing the same question on both exchanges; 31 produced matches |
| Matching | Outcome names normalised and matched one to one; ambiguous matches dropped |
| Matched outcomes | 351 |
| Two-sided on both exchanges | 151 |
| Liquid sample | 91: both spreads of 5¢ or less, price between 3% and 97% |
| Price measure | Mid-point of the best bid and best ask on each exchange |
For each event we fetched both order books one after the other, within seconds. Outcomes were matched by name after normalising spelling, accents and party labels, so "Ken Paxton (R)" on Polymarket matched "Ken Paxton" on Kalshi and "Miami (FL)" matched "Miami Hurricanes." Where a match was ambiguous, such as Miami (Ohio) against Miami (Florida), we dropped it.
The events cover the Senate and House control markets, eleven Senate races, three governor races, the 2028 presidential election, the December Fed decision, a 2026 US recession, and sports futures and awards: the NBA, NFL and college football champions, the Heisman, NFL MVP, NBA Rookie of the Year, the 2027 World Series, MLB's Cy Young and MVP races, the NHL's Hart Trophy and the UFC heavyweight title.
Polymarket here means the international, crypto-based site, not the separate CFTC-regulated Polymarket US app, which has its own order book. The full list of matched outcomes with both books is in the open data file.
How big are the gaps?
Source: Our analysis; 91 liquid matched outcomes, 30 October 2026
Show the dataHide the data
| Value | |
|---|---|
| 0–0.5 | 24 |
| 0.5–1 | 26 |
| 1–2 | 22 |
| 2–3 | 11 |
| 3–5 | 7 |
| Over 5 | 1 |
Half of the liquid outcomes, 50 of 91, were within 1 point. Most of the rest were within 3. The single outlier was the 2028 winning party market: Polymarket priced the Republicans at 34.5% and Kalshi at 40.5%, a 6-point gap in a market that will not settle for two years.
Which exchange is cheaper?
Source: Our analysis, 30 October 2026
Show the dataHide the data
| Value | |
|---|---|
| Kalshi | 36 |
| Polymarket | 31 |
| Within 0.5 point | 24 |
Neither exchange is consistently cheaper. "Yes" was cheaper on Kalshi in 36 outcomes and on Polymarket in 31. That is close to an even split, so there is no rule of thumb such as "always buy on Kalshi." The cheaper side depends on the market, and it changes from day to day. Our Polymarket vs Kalshi comparison page shows the live gaps.
Spreads were similar too. Both exchanges had a median spread of 1¢ in the liquid sample. Kalshi's book was tighter more often, 33 times against 22, and its average spread was lower (1.01¢ against 1.34¢).
Where the gaps are largest
Source: Our analysis, 30 October 2026
Show the dataHide the data
| Value | |
|---|---|
| Most active third | 1.00 pt |
| Middle third | 0.55 pt |
| Least active third | 1.65 pt |
The least active third of outcomes had the widest gaps, a median of 1.65 points and an average of 1.95. Busy markets attract traders who watch both exchanges and close gaps quickly.
By price level, the smallest gaps were in long shots priced at 3–10% (median 0.55 points), and the largest in the 10–30% range (median 2.0 points). By category, the median gap was 1.0 point in politics, economics and sports alike; sports had a slightly lower average.
Republicans cost more on Kalshi
The most consistent pattern was political. In the 15 races and control markets where we could compare the Republican side, Kalshi priced the Republican higher in 11, lower in 1 and about the same in 3:
| Market | Republican side | Polymarket | Kalshi | Kalshi minus Polymarket |
|---|---|---|---|---|
| 2028 winning party | Republican | 34.5% | 40.5% | +6.0 |
| Senate control | Republican | 27.5% | 30.5% | +3.0 |
| Kansas Senate | Roger Marshall | 45.5% | 48.5% | +3.0 |
| Michigan Senate | Mike Rogers | 22.5% | 25.5% | +3.0 |
| Ohio governor | Vivek Ramaswamy | 18.5% | 20.5% | +2.0 |
| New Hampshire Senate | John E. Sununu | 7.0% | 8.3% | +1.2 |
| Texas Senate | Ken Paxton | 28.5% | 29.5% | +1.0 |
| Maine Senate | Susan Collins | 31.5% | 32.5% | +1.0 |
| Ohio Senate | Jon Husted | 29.5% | 30.5% | +1.0 |
| House control | Republican | 6.5% | 7.4% | +0.9 |
| California governor | Steve Hilton | 2.5% | 3.0% | +0.6 |
| Georgia Senate | Mike Collins | 3.0% | 3.3% | +0.2 |
| Georgia governor | Rick Jackson | 34.5% | 34.5% | 0.0 |
| Iowa Senate | Ashley Hinson | 55.5% | 55.5% | 0.0 |
| Nebraska Senate | Pete Ricketts | 67.5% | 66.5% | −1.0 |
Mid-prices, 30 October 2026, about 15:48 UTC.
On average the Republican side was 1.5 points dearer on Kalshi. The likely explanation is the customer base: Kalshi's users are US residents trading in dollars, while Polymarket's international site serves traders outside the US. The gap does not tell you which exchange is right, only that the two crowds lean differently. Our Senate control odds and midterm predictions cover the races themselves.
Crossed books: free money?
When one exchange's best bid for "Yes" is higher than the other exchange's best ask, you can buy "Yes" where it is cheap and buy "No" on the other exchange, and the two together cost less than the $1 one of them will pay. That is a crossed book. We found 39 in the liquid sample before fees. The largest:
| Market | Polymarket bid/ask | Kalshi bid/ask | Edge before fees |
|---|---|---|---|
| 2028 winning party: Republican | 34¢ / 35¢ | 40¢ / 41¢ | 5.0¢ |
| 2028 winning party: Democratic | 64¢ / 65¢ | 60¢ / 61¢ | 3.0¢ |
| US recession in 2026 | 6¢ / 8¢ | 2¢ / 3¢ | 3.0¢ |
| 2028 president: Alexandria Ocasio-Cortez | 12.4¢ / 12.5¢ | 9.8¢ / 9.9¢ | 2.5¢ |
| 2028 president: Jon Ossoff | 9.4¢ / 9.5¢ | 12¢ / 13¢ | 2.5¢ |
| Senate control: Republican | 27¢ / 28¢ | 30¢ / 31¢ | 2.0¢ |
| Kansas Senate: Roger Marshall | 45¢ / 46¢ | 48¢ / 49¢ | 2.0¢ |
| Michigan Senate: Mike Rogers | 22¢ / 23¢ | 25¢ / 26¢ | 2.0¢ |
Edge = what one pair of contracts pays ($1) minus what it costs, before fees.
In practice most of that edge disappears:
- Fees. After Kalshi's taker fee on the Kalshi leg alone, only 16 of the 39 crossed books stayed positive. Polymarket's international site also charges taker fees in many categories, which would cut the number further. Our guides to Kalshi fees and Polymarket fees vs Kalshi fees have the formulas.
- Size. We measured the best prices, not how many contracts were available at them. The top of the book is often only a few hundred dollars deep.
- Time. Money in the 2028 markets is locked up for two years. A 3¢ edge over two years is a poor return.
- Rules. The two exchanges' contracts are similar, not identical. Settlement sources, deadlines and edge cases can differ; read both rulebooks.
- Moving money. You need funded accounts on both exchanges, one in US dollars and one in crypto on the international site, which US residents cannot use.
Our guide to prediction market arbitrage goes through a worked example.
What about the illiquid outcomes?
Of the 351 matched outcomes, 200 did not have a two-sided quote on both exchanges, mostly long shots with no bids. Another 18 had a two-sided quote but a spread wider than 5¢ on at least one exchange, usually Polymarket. The 2027 World Series market was the clearest case: Polymarket's books had spreads of 17¢ to 75¢, so its mid-prices meant little, and comparing them with Kalshi's produced gaps of up to 35 points. We excluded those from the main results. Thin markets drift apart because few traders watch both, and because the cost of trading them is high: see our study of bid-ask spreads.
What it means for traders
- Check both prices before you trade. A 1–3 point difference is common, and on a 30¢ contract 2 points is about 7% of the price.
- Use limit orders in thin markets, and compare the ask, not the mid-price, with the other exchange.
- Do not chase arbitrage blindly. After fees, depth and capital costs, most crossed books are not worth it for retail traders.
- Watch the political skew. If you want the Democratic side of a race, Kalshi was often the cheaper place to buy it on our snapshot; for the Republican side, often Polymarket.
Limitations
This is a single snapshot taken five days before the midterms, when election markets were unusually active. Gaps change by the minute. The two order books were fetched a few seconds apart, not at the same instant. We compared mid-prices; executable prices are the bid and ask. Polymarket here is the international site, which US residents cannot use. We will repeat the study to see which patterns hold.
Data
- Matched outcomes with both order books (CSV): snapshot time, category, event, outcome on each exchange, best bid and ask, 24-hour and total volume, and whether the outcome is in the liquid sample.
Related reading
- Kalshi vs Polymarket: full comparison
- Prediction market arbitrage
- Bid-ask spreads on Kalshi and Polymarket
- What do people bet on? Volume by category
- How accurate are prediction markets?
This study uses public market data and is not investment advice. Prices change continuously.
