Funding rate arbitrage scanner
Gross BTC and ETH funding spreads from stored current-rate rows. Each pair passes an open-interest filter; execution capacity and net costs remain unmeasured.
18.7%
widest BTC spread, annualized · Apex Omni vs Kraken
38.2%
widest ETH spread, annualized · dYdX vs BloFin
240
pairs passing the OI filter in the latest stored asset sweeps
Open interest (OI) must reach $1,000,000 on each side. OI does not measure available order size. Net results after fees, spread, slippage, and borrowing remain incomplete.
BTC perpetual funding divergence
BTC — pairs passing the OI filter (12 of 120)
| Long leg | Short leg | Long rate 1h | Short rate 1h | Spread ann. | Min OI (USD) | Older source |
|---|---|---|---|---|---|---|
| Apex Omni | Kraken | 0.00023% | 0.00237% | 18.7% | $130.5m | 10-24 14:37 UTC |
| Phemex | Kraken | 0.00031% | 0.00237% | 18.1% | $185.9m | 10-24 14:37 UTC |
| Apex Omni | GMX | 0.00023% | 0.00199% | 15.3% | $20.9m | 10-24 14:37 UTC |
| Bybit | Kraken | 0.00063% | 0.00237% | 15.2% | $225.4m | 10-24 14:37 UTC |
| Phemex | GMX | 0.00031% | 0.00199% | 14.7% | $20.9m | 10-24 14:37 UTC |
| Apex Omni | BloFin | 0.00023% | 0.00165% | 12.4% | $130.5m | 10-24 14:37 UTC |
| Bybit | GMX | 0.00063% | 0.00199% | 11.8% | $20.9m | 10-24 14:37 UTC |
| Phemex | BloFin | 0.00031% | 0.00165% | 11.8% | $185.9m | 10-24 14:37 UTC |
| Bybit | BloFin | 0.00063% | 0.00165% | 8.9% | $324.1m | 10-24 14:37 UTC |
| Deribit | Crypto.com | 0.00088% | 0.00130% | 3.6% | $158.1m | 10-24 14:37 UTC |
| Deribit | Bitget | 0.00088% | 0.00125% | 3.2% | $158.1m | 10-24 14:37 UTC |
| Deribit | KuCoin | 0.00088% | 0.00125% | 3.2% | $158.1m | 10-24 14:37 UTC |
Both markets report at least $1,000,000 open interest. Rates use a common one-hour period. The gross spread excludes fees, spread, slippage, and borrowing costs; net results remain incomplete. Rows sort by spread, with at most 3 rows per venue. The OI filter does not establish fill quality.
BTC — venue matrix (17 venues with reported open interest)
Sweep 10-24 14:37 UTC| Venue | Type | Normalized 1h | Annualized | Open interest (USD) | Interval | Passes OI filter |
|---|---|---|---|---|---|---|
| Kraken | cex | 0.00237% | 20.8% | $225.4m | 1h | Yes |
| GMX | dex | 0.00199% | 17.4% | $20.9m | 1h* | Yes |
| BloFin | cex | 0.00165% | 14.5% | $324.1m | 8h | Yes |
| Crypto.com | cex | 0.00130% | 11.4% | $498.3m | 1h* | Yes |
| Bitget | cex | 0.00125% | 11.0% | $2.95b | 8h | Yes |
| KuCoin | cex | 0.00125% | 11.0% | $737.1m | 8h | Yes |
| Hyperliquid | dex | 0.00125% | 11.0% | $3.48b | 1h* | Yes |
| Orderly Network | dex | 0.00125% | 11.0% | $11.2m | 8h | Yes |
| Bluefin | dex | 0.00125% | 11.0% | $77.5k | 1h* | No |
| GRVT | dex | 0.00125% | 11.0% | $202.8m | 8h* | Yes |
| Lighter | dex | 0.00120% | 10.5% | $167.8m | 1h* | Yes |
| dYdX | dex | 0.00118% | 10.3% | $19.0m | 1h* | Yes |
| Paradex | dex | 0.00115% | 10.0% | $6.7m | 8h* | Yes |
| Deribit | cex | 0.00088% | 7.7% | $158.1m | 8h* | Yes |
| Bybit | cex | 0.00063% | 5.6% | $5.14b | 8h | Yes |
| Phemex | cex | 0.00031% | 2.7% | $185.9m | 8h | Yes |
| Apex Omni | dex | 0.00023% | 2.1% | $130.5m | 1h* | Yes |
8 additional BTC venues have a current rate but no open-interest value in this sweep. The page cannot test those rows against the $1,000,000 floor. An asterisk identifies an explicitly assumed interval. Some direct adapters use fixed schedules without this marker. The methodology lists each interval rule.
ETH perpetual funding divergence
ETH — pairs passing the OI filter (12 of 120)
Older source 10-24 14:37 UTC| Long leg | Short leg | Long rate 1h | Short rate 1h | Spread ann. | Min OI (USD) |
|---|---|---|---|---|---|
| dYdX | BloFin | -0.00219% | 0.00216% | 38.2% | $25.3m |
| dYdX | Kraken | -0.00219% | 0.00184% | 35.3% | $25.3m |
| dYdX | Bitget | -0.00219% | 0.00125% | 30.2% | $25.3m |
| Lighter | BloFin | -0.00010% | 0.00216% | 19.8% | $82.8m |
| Lighter | Kraken | -0.00010% | 0.00184% | 17.0% | $67.6m |
| GRVT | BloFin | 0.00025% | 0.00216% | 16.8% | $136.6m |
| GRVT | Kraken | 0.00025% | 0.00184% | 13.9% | $67.6m |
| Lighter | Bitget | -0.00010% | 0.00125% | 11.8% | $82.8m |
| GRVT | Bitget | 0.00025% | 0.00125% | 8.8% | $136.6m |
| GMX | KuCoin | 0.00037% | 0.00125% | 7.7% | $15.4m |
| GMX | Orderly Network | 0.00037% | 0.00125% | 7.7% | $15.4m |
| GMX | Crypto.com | 0.00037% | 0.00118% | 7.1% | $15.4m |
Both markets report at least $1,000,000 open interest. Rates use a common one-hour period. The gross spread excludes fees, spread, slippage, and borrowing costs; net results remain incomplete. Rows sort by spread, with at most 3 rows per venue. The OI filter does not establish fill quality.
ETH — venue matrix (17 venues with reported open interest)
Sweep 10-24 14:37 UTC| Venue | Type | Normalized 1h | Annualized | Open interest (USD) | Interval | Passes OI filter |
|---|---|---|---|---|---|---|
| BloFin | cex | 0.00216% | 19.0% | $220.7m | 8h | Yes |
| Kraken | cex | 0.00184% | 16.1% | $67.6m | 1h | Yes |
| Bitget | cex | 0.00125% | 11.0% | $1.87b | 8h | Yes |
| KuCoin | cex | 0.00125% | 11.0% | $426.2m | 8h | Yes |
| Orderly Network | dex | 0.00125% | 11.0% | $24.3m | 8h | Yes |
| Bluefin | dex | 0.00125% | 11.0% | $43.3k | 1h* | No |
| Crypto.com | cex | 0.00118% | 10.3% | $414.0m | 1h* | Yes |
| Hyperliquid | dex | 0.00103% | 9.0% | $2.59b | 1h* | Yes |
| Deribit | cex | 0.00096% | 8.4% | $52.8m | 8h* | Yes |
| Bybit | cex | 0.00086% | 7.5% | $1.88b | 8h | Yes |
| Phemex | cex | 0.00078% | 6.8% | $246.5m | 8h | Yes |
| Paradex | dex | 0.00055% | 4.8% | $2.6m | 8h* | Yes |
| Apex Omni | dex | 0.00045% | 4.0% | $4.6m | 1h* | Yes |
| GMX | dex | 0.00037% | 3.2% | $15.4m | 1h* | Yes |
| GRVT | dex | 0.00025% | 2.2% | $136.6m | 8h* | Yes |
| Lighter | dex | -0.00010% | -0.9% | $82.8m | 1h* | Yes |
| dYdX | dex | -0.00219% | -19.2% | $25.3m | 1h* | Yes |
8 additional ETH venues have a current rate but no open-interest value in this sweep. The page cannot test those rows against the $1,000,000 floor. An asterisk identifies an explicitly assumed interval. Some direct adapters use fixed schedules without this marker. The methodology lists each interval rule.
How the spread is measured
Funding divergence is the difference between two venues' rates for the same asset. PerpFinder normalizes current rates to one hour and subtracts the lower rate from the higher rate. Both rows come from one collection run; their source timestamps can differ.
The displayed pair places the long position at the lower-rate venue and the short position at the higher-rate venue. Opposite positions can reduce exposure to asset price changes, but basis changes and separate liquidation rules remain relevant.
The annualized spread assumes the hourly difference persists for 8,760 hours. Minimum OI means the smaller reported open interest of the two positions' markets. It does not set an order-size limit or establish available liquidity.
Largest spread after the OI filter · version 4
established- BTC: 30 stored days, 2842 sweeps · widest 96.1% on 2026-09-25 · typical 22.2%
- ETH: 30 stored days, 2842 sweeps · widest 217.0% on 2026-10-13 · typical 25.7%
Each point shows the largest hourly spread among eligible pairs and sweeps for one UTC day, extrapolated to one year. The reader selects methodology version 4 within the latest 30 stored day files. The returned series can be shorter than that window. Earlier method versions do not enter these points. “Typical” is the median of daily median spreads. Rates and costs can change; these gross observations do not establish future returns.
Worked example — BTC, $100,000 per leg
Assume $100,000 of constant notional value on each side of this observed BTC pair. At unchanged rates, the gross funding difference totals $359 over 7 days. The long position uses Apex Omni; the short position uses Kraken.
The calculation is 0.00214% per hour × 168 hours × $100,000. The pair timestamp is 10-24 14:37 UTC, the older source time of its two rows. Rates can change before settlement. This hypothetical result excludes costs and does not forecast income.
Net opportunity after costs: incomplete
Net results require entry and exit fees, bid-ask spread, slippage, and any borrowing cost for both positions. Future funding rates, holding duration, and settlement intervals also affect the result. This page does not combine all these inputs, so the net result remains incomplete.
The cost comparison tool supplies separate execution-cost observations. Those observations do not establish a net return for the funding pairs shown here.
Risks in a funding pair
Funding arbitrage is not risk-free. Rates can change before settlement, and liquidation can close one position while the other remains open. Venue failures, access restrictions, and withdrawal limits can affect either account.
Each venue requires its own margin. Changes in funding intervals can also change normalized comparisons. The methodology identifies observed intervals, fixed assumptions, and timestamp limits.
The funding overview separates available current and predicted rates. The history page links funding records and shows price-basis charts. History requests can return fewer than 30 days when the selected method version has a shorter record.
Questions about funding divergence
What is cross-venue funding divergence?
It is the difference between two venues’ current funding rates for the same asset.
PerpFinder converts both rates to one hour before subtraction. The rows come from one collection run, but their source timestamps can differ.
What does an annualized funding spread mean?
The annualized figure extends an observed hourly difference across 8,760 hours at an unchanged rate.
It permits comparisons on one scale. Funding can change before settlement, so the figure does not forecast a yearly return.
What does the open-interest filter check?
Both markets must report at least $1,000,000 of open interest in the same collection run.
The filter does not measure order-book capacity, fill quality, or an available trade size.
How often does collection run?
The collection target is every 15 minutes.
Tables use the latest stored asset sweeps and their timestamps. The spread chart selects the current method version from up to 30 stored day files. Missing observations can shorten the series.
Why are net results after costs incomplete?
The page does not combine entry and exit fees, spread, slippage, borrowing costs, and the actual funding path for both positions.
The displayed spreads therefore remain gross, and net results remain incomplete.