Why is Bitcoin price different on exchanges?
Coinbase Bitcoin was a median 0.03% above Binance over 62,342 hours since 2019. See the gap by year and why currency and contract type matter.
Bitcoin's price differs between exchanges because each exchange records its own trades, and the two prices may be in different currencies or for different products. Across 62,342 paired hourly closes from September 8, 2019, to October 20, 2026, Coinbase BTC-USD spot was a median 0.03% above Binance BTCUSDT perpetual. 90% of the hourly gaps fell between −0.12% and +0.15%, and the gap went in both directions every year.
That median is about 3 cents for every $100 of Bitcoin. The median is the middle gap after sorting all of them from lowest to highest. This one number mixes three differences at once: exchange, currency and contract type.
For why this gap matters when you search for similar charts, see Why chart pattern matching should stay on one exchange.
Are both screens showing the same kind of Bitcoin market?
Start with the full market name. BTC means Bitcoin, but the rest of the name matters too.
On Coinbase BTC-USD spot, the price is in US dollars. Spot means trading the coin itself. On Binance BTCUSDT perpetual, the price is in USDT, a token designed to track the dollar. A perpetual, often called a perp, is a futures contract tied to Bitcoin's price that never expires.
So this comparison changes more than the exchange. It also changes what is traded and the unit the price is written in. A USDT price is not automatically a dollar price.
Each exchange also has its own buyers, sellers and trades. The last price on one screen is the last trade there. It can differ from the last trade somewhere else.
How large was the Coinbase–Binance price gap?
We compared hourly bars. A bar records the opening, highest, lowest and closing price during a set period. We paired the closing prices that had the same UTC time label. UTC is a shared world time standard.
For each pair, we divided the Coinbase close by the Binance close, subtracted one, and wrote the result as a percentage. A positive gap means the number on Coinbase was higher. We did not convert USDT into dollars first.
Every row covers Bitcoin on Coinbase BTC-USD spot and Binance BTCUSDT perpetual, using hourly closes. The first and last years are partial.
| UTC close year | Paired hourly closes | Median gap | 5th-percentile gap | 95th-percentile gap |
|---|---|---|---|---|
| 2019, from September 8 | 2,741 | +0.06% | −0.20% | +0.30% |
| 2020 | 8,781 | +0.03% | −0.23% | +0.25% |
| 2021 | 8,760 | +0.01% | −0.15% | +0.15% |
| 2022 | 8,760 | +0.05% | −0.10% | +0.12% |
| 2023 | 8,757 | +0.04% | −0.08% | +0.16% |
| 2024 | 8,784 | +0.01% | −0.12% | +0.10% |
| 2025 | 8,755 | +0.05% | −0.04% | +0.11% |
| 2026, through October 20 | 7,004 | −0.01% | −0.11% | +0.08% |
| Whole study period | 62,342 | +0.03% | −0.12% | +0.15% |
The percentiles are cutoffs in the sorted gaps. About 5% of the gaps fall below the lower cutoff, and about 5% fall above the upper one. They are not the smallest and largest gaps; some hours went further.
Was Coinbase always more expensive?
No. In every year, the lower cutoff was negative and the upper cutoff was positive. So the gap went both ways each year.
The middle gap changed direction too. In 2025, the median across 8,755 paired hours was +0.05%. In 2026 through October 20, it was −0.01% across 7,004 pairs.
The spread between the cutoffs changed as well. In 2020, the cutoffs were −0.23% and +0.25%. In 2025, they were −0.04% and +0.11%. The later range was narrower, but this test does not tell us why.
What can cause the prices to differ?
The measured gap mixes several things that this test cannot separate.
Currency. USDT is designed to track the dollar, but it is a different unit. Any difference between their values shows up in this gap, because we did not convert.
The contract. A perp can trade above or below spot. That difference is called the basis. Perps also have funding: regular payments that traders on one side of the contract make to the other side. Our gap does not include those payments. Our guide to funding rate history explains that separate cost.
Trade timing. The same hourly label does not mean both last trades happened at the same instant. Exchanges can also build their closing prices in slightly different ways.
These are possible parts of the answer. The table cannot say how much each one added, or which one caused a given year's change.
Does the measured gap show a trading profit?
No. Arbitrage means trying to profit from price differences between related markets. We did not test that.
A closing price is a record of one past trade. It does not show the bid (the best price a buyer offers) or the ask (the best price a seller wants) at that moment.
The table also leaves out fees, funding, limits on moving money between exchanges, and slippage. Slippage is the gap between the price you expect and the price you actually get. Hourly closes also do not show how long a gap lasted inside the hour. The numbers alone do not show a profit after costs.
How can you use these numbers in your own research?
Use the yearly cutoffs as a "normal background" check. If Coinbase BTC-USD spot and Binance BTCUSDT perpetual differ by an amount inside that year's 5th-to-95th range (for 2025, −0.04% to +0.11%), that gap was ordinary in past hours. It is not, by itself, a sign that something unusual is happening.
Second, keep each study on one market. If you mark a price level, a high or a pattern on one exchange's chart, check it on that same exchange's chart. A small gap between markets can make a price look like it crossed a level on one screen but not the other. If you need price history from several exchanges, our comparison of crypto data APIs shows which tools carry which markets.
How we measured this
We used public Coinbase BTC-USD spot and Binance BTCUSDT perpetual hourly price histories. We paired completed hourly closes with the same UTC time label, from September 8, 2019, through October 20, 2026, and left out hours where either market had no price. We divided each Coinbase close by its Binance close, subtracted one, and found the median and the 5th and 95th percentiles for the whole period and for each year. We did not convert USDT to USD or include fees, funding, slippage or transfer limits. Nearby hours share price history, so the 62,342 pairs are not 62,342 separate events, and the results describe only these two markets.
Why does Comparand keep chart searches on the same exchange?
Because the market name matters. Coinbase Bitcoin spot and Binance Bitcoin perps have different price histories, even though both track Bitcoin.
Comparand searches a coin's own past on the same exchange and chart timeframe. A Binance Bitcoin search stays within Binance Bitcoin history. Each result shows its own price path and what happened next.
These are past cases, not a forecast. Read what a chart analog means for how the search works, or why past moves are not forecasts for its limits.
Try this in Comparand
Ask an AI assistant connected to Comparand:
"Which exchanges and kinds of data can you get? Find Coinbase BTC-USD spot and Binance BTCUSDT perpetual. Get completed hourly bars for the same past day on both, and work out the hourly gap. If individual trades are available, fetch them for one hour of that day too."
What comes back:
- venue_coverage: the exchanges served, with the kinds of data and time intervals each one has.
- venue_markets: each exchange's list of markets, so you can pick the exact Bitcoin symbol and market type.
- venue_bars: the chosen market's open, high, low, close and volume for completed bars in that period.
- venue_trades: individual trades for the chosen symbol on exchanges that publish them, at most 24 hours per request.
These four tools are free on every key. Your assistant can then work out the gap from the two sets of bars. See the connection docs, or try 3 coins free to explore past charts.
The short version
- Check the exchange, the currency and whether the price is spot or a perp.
- From 2019 to 2026, Coinbase BTC-USD spot was a median 0.03% above Binance BTCUSDT perp, and 90% of hourly gaps fell between −0.12% and +0.15%.
- The gap went both ways every year, and its size changed by year.
- Closing-price gaps alone do not show a trading profit.
- Comparand searches a coin's own past on the same exchange and timeframe.
Questions people ask
Why do Coinbase and Binance show different Bitcoin prices?
Each exchange has its own buyers, sellers and trades, so each records its own last price. The two prices can also be in different currencies or for different products, such as Coinbase BTC-USD spot and Binance BTCUSDT perpetual.
How big is the Bitcoin price gap between Coinbase and Binance?
In 62,342 paired hourly closes from September 8, 2019, to October 20, 2026, Coinbase BTC-USD spot was a median 0.03% above Binance BTCUSDT perpetual. 90% of the hourly gaps fell between -0.12% and +0.15%, and the gap went both ways every year.
Is BTC-USD the same as BTCUSDT?
Both track Bitcoin, but USD means US dollars and USDT is a token designed to track the dollar. The name alone also does not tell you whether the market is Bitcoin itself (spot) or a contract tied to its price (a perp).
Which Bitcoin price is the real one?
There is no single official Bitcoin price. Each exchange's price is a real last trade on that one market. So when you see a Bitcoin price, check which exchange and which market it comes from.
Does a Bitcoin price gap between exchanges mean easy profit?
No. Closing prices leave out fees, funding, slippage and the cost and delay of moving money between exchanges. They also do not show the prices you could actually trade at in that moment.