One Concept, Four Field Names, Three Numbers: The Funding-Rate Ceiling on the Same BTC Perp

Disclosure: this article was researched and drafted with AI assistance, using public exchange APIs and documentation, and reviewed for accuracy against the sources cited below.

This is not financial advice. All numbers below are parameter values read from public exchange APIs at a specific moment in time. Exchanges can and do change these values without notice. Always verify against the official sources listed at the end.

The number everyone quotes doesn't exist

Ask five traders what the maximum funding rate on a Bitcoin perpetual is and you'll get a confident shrug. There is no industry-wide ceiling. Funding-rate caps are per-exchange, per-contract parameters — and even for the single most liquid contract in crypto, the BTC/USDT perpetual, the four venues examined here publish three different answers under four different field names.

I pulled all of them live, without an API key, from official public endpoints on October 4, 2026 at 13:37 UTC. Here is what the same concept looks like at four exchanges.

The four readings, side by side

Venue	Field name(s)	BTC perp cap	ETH perp cap	Settlement interval
Gate	funding_rate_limit	±0.30%	±0.30%	28800 s (8 h)
KuCoin	fundingRateCap / fundingRateFloor	±0.30%	±0.375%	28800000 ms (8 h)
Bybit	upperFundingRate / lowerFundingRate (specs) + fundingCap (tickers)	±0.333%	±0.333%	480 min (8 h)
OKX	maxFundingRate / minFundingRate	±0.375%	±0.75%	8 h

Three numerical answers for the identical underlying exposure: 0.30%, 0.333%, and 0.375%. The spread between the lowest and highest cap is 25% — small in absolute terms, but it defines how far an overheated or capitulating market can push the cost of holding a position before the ceiling bites.

Live funding values at the same moment, for context (these change every interval and are not the point of this article): Gate BTC perp −0.0017%/8h, Bybit +0.006547%/8h, OKX +0.0027116%/8h, KuCoin −0.0002%/8h.

Four names for one concept

If you are writing any tooling that compares funding terms across venues, the field names alone will humble you:

• Gate exposes funding_rate_limit in its futures contracts model. The field's meaning is confirmed in Gate's own public API documentation changelog, which added the field "to indicate funding rate cap value."
• Bybit publishes the same parameter twice under two names: upperFundingRate/lowerFundingRate in the instruments-specification endpoint, and fundingCap in the tickers endpoint. On October 4 both endpoints returned 0.00333 for BTCUSDT — a useful internal cross-check you can reproduce yourself.
• OKX calls it maxFundingRate/minFundingRate on its funding-rate endpoint.
• KuCoin uses fundingRateCap/fundingRateFloor — and note the data types differ too: KuCoin returns numbers, while Gate, Bybit, and OKX return strings.

One concept, four spellings, two data types. There is no standard.

The ETH side is where the philosophies split

The BTC caps differ, but the more interesting divergence is on the second-most liquid contract. At the same moment:

• OKX sets ETH's cap at ±0.75% — exactly double its own BTC cap of ±0.375%.
• KuCoin sets ETH at ±0.375% — exactly double its own BTC cap of ±0.30%.
• Gate (±0.30%) and Bybit (±0.333%) give ETH the same ceiling as BTC.

So two venues deliberately price volatility headroom by asset, and two treat majors as one class. Neither approach is right or wrong — but it means "the cap on this exchange" is not a single number you can memorize per venue. It is a matrix: venue × contract.

Why the ceiling matters even if you never see it

Most funding intervals settle near zero — a few basis points every eight hours. The caps are invisible in calm markets. They matter in the tails: when a crowded position unwinds, funding can pin at the cap for consecutive intervals, and that pinned value is your holding cost, compounding every settlement period. A position on the venue with a 0.30% cap faces a structurally different worst-case carrying cost than the same position where the cap is 0.375% — and, as previous measurements in this series have shown, some exchanges settle smaller contracts hourly or every four hours, where the same cap applied more often multiplies the effect.

Two caveats that deserve repetition: the caps are parameters exchanges can adjust at will, and the interval length is a separate field you must read before comparing any two caps. "0.333% every 8 hours" and "2.5% every 1 hour" are not comparable numbers.

How to verify every number in this article

No login or API key is needed for any of it:

• Gate: GET https://api.gateio.ws/api/v4/futures/usdt/contracts/BTC_USDT (and ETH_USDT) — read funding_rate_limit.
• Bybit: GET https://api.bybit.com/v5/market/instruments-info?category=linear&symbol=BTCUSDT — read upperFundingRate; cross-check GET https://api.bybit.com/v5/market/tickers?category=linear&symbol=BTCUSDT — read fundingCap.
• OKX: GET https://www.okx.com/api/v5/public/funding-rate?instId=BTC-USDT-SWAP (and ETH-USDT-SWAP) — read maxFundingRate.
• KuCoin: GET https://api-futures.kucoin.com/api/v1/contracts/XBTUSDTM (and ETHUSDTM) — read fundingRateCap.

Snapshot taken 2026-10-04 13:37 UTC. If you re-run these calls next week and the numbers moved, that is the exchange exercising the very parameter this article describes — not an error in the measurements.

---

Not financial advice. This article describes publicly published contract parameters. It makes no recommendation to trade, hold, or move positions on any venue named. Perpetual futures involve substantial risk of loss. Funding rates, caps, and settlement intervals change without notice.

#ad — affiliate disclosure: no sponsored or tracked links appear in this article; no exchange named here has paid for inclusion. If this publication adds such links in the future, they will be marked the same way, before the link.

Verify official sources: always confirm current terms directly on each exchange's official API documentation and fee pages before relying on any number — exchanges revise these parameters at their discretion.