Base rates
All fees are expressed in basis points (bps). 1 bps = 0.01%.“Cross” means taker (crossing the spread). “Add” means maker (adding liquidity to the book).
These are the rates for users with no volume history, no staking, and no referral. Every discount mechanism below reduces from these base rates.
VIP tiers
Volume-based fee reductions. Tier qualification is based on 14-day rolling notional traded volume (NTL). Spot volume may count at 2x toward tier qualification when theweighSpotVolumeDouble flag is active.
Fee tiers are reassessed automatically based on rolling 14-day volume. Users can also trigger a manual reassessment via the
reassessFees action.
Market maker rebates
Separate from VIP tiers, the MM rebate program rewards users who contribute a high fraction of maker (add) volume. Qualification is based on maker fraction — the percentage of your total volume that is maker.Negative maker rates mean you get paid to provide liquidity. MM rebates stack with VIP tiers. A VIP 3 user with 3%+ maker fraction would receive -0.003% maker instead of paying +0.004%.
HYPE staking discounts
Staking HYPE tokens reduces trading fees. Discount tiers are based on basis points of the max HYPE supply staked:
Staking discounts apply as a percentage reduction on your effective fee rate. A 40% staking discount on a 4.5 bps taker rate brings it down to 2.7 bps.
To activate staking discounts, you must link your staking address to your trading address via the
linkStakingUser action. The per-user API returns stakingLink, activeStakingDiscount, and the full stakingDiscountTiers table.
Referral discount
Users who set a referrer receive a 4% discount on all trading fees.- Referred users get 4% off their effective rate
- Referrers receive a fraction of that 4% as revenue
- Setting a referrer is a one-time action (
setReferrer) - Referrer registration uses
RegisterReferrerAction
Staking discounts and referral discounts stack multiplicatively. Maximum combined discount: 4% referral + 40% staking = up to ~42.4% total fee reduction.
Builder fees
Builder fees are additional fees charged on top of protocol fees by frontends, bots, or SDKs that route orders through HyperLiquid. Key mechanics:- Users must explicitly approve each builder via
approveBuilderFee - The builder address and maximum fee capacity are embedded in each order
- Builder fees are collected separately and tracked in
CollectedBuilderFees - Builders can be any address (frontend operator, bot developer, SDK author)
Deployer fee share
HIP-3 token deployers earn a share of all trading fees generated by their deployed assets.
Deployers can update their fee recipient and fee scale, though scale changes are rate-limited by
last_deployer_fee_scale_change_time. Sub-deployers can also be designated to receive fee shares.
This creates an economic flywheel: deployers are incentivized to create and promote active markets because they earn ongoing revenue from trading activity.
HIP-4 prediction market deployers also earn fee share on outcome token trading. See HIP-4 for deployment details.
Fee trials
Fee trials let users temporarily access lower fee tiers without meeting the volume requirements.
Starting a trial requires escrowing USDC. During the trial period, an override fee schedule applies. After expiry, there is a cooldown period before the next trial can begin. Fee trials can be globally disabled via the
trials_disabled flag.
Fee tier override states
The protocol can override a user’s fee tier in two ways:
This serves as a penalty mechanism for users who violate protocol rules.
HIP-4 outcome fees
Prediction market (HIP-4) outcomes use spot fee rails:- Opening a position is free (no fee on SplitOutcome)
- Closing on the orderbook charges spot taker/maker rates
- Settlement charges fees in the outcome token, not USDC
- A governance hook (
SetOutcomeFeeScale) can adjust outcome fees independently
Aligned quote token scaling
Different quote tokens can have different fee scales:
This allows the protocol to incentivize or disincentivize trading in specific settlement tokens.
Querying fees via API
Request weight system
HyperLiquid implements a request weight system that governs API rate limits. Every API request consumes a certain amount of request weight from the user’s allocation.Expanding request weight
Users can purchase additional request weight through thereserveRequestWeight action at a cost of 0.0005 USDC per request weight unit. This provides a self-service mechanism to scale API access beyond the default allocation.
Rate limit queries
Users can check their current rate limit status via theuserRateLimit info endpoint, which returns:
- Current request weight consumption
- Remaining allocation
- Reset timing
Request weight is separate from WebSocket subscription limits. A user on a higher fee tier automatically receives higher default request weight allocations.
Fee schedule internals
The protocol maintains aFeeSchedule structure that encapsulates all fee configuration:
Fee trials
Fee trials allow users to temporarily test a lower fee tier:- User initiates via
startFeeTrialaction - An escrow is placed for the trial period
- During the trial, the user pays the lower tier’s rates
- After the trial ends, the escrow is returned and fees revert
- A cooldown period prevents immediate re-trials
Fee reassessment
ThereassessFees action triggers a recalculation of a user’s fee tier based on their current 14-day rolling volume. This is useful after large trading periods to ensure the correct tier is applied without waiting for the automatic reassessment cycle.
Override states
Fee tiers can be overridden through several mechanisms:
When multiple overrides apply, they stack multiplicatively (not additively). For example, a 4% referral discount combined with a 40% staking discount yields approximately 42.4% total discount, not 44%.