Hyperliquid TWAP Orders: 7-Day Windows, Dynamic Intervals & $100 Minimums

Hyperliquid

TWAP Orders Just Got a Lot More Useful

A TWAP order splits one large trade into a run of small ones and feeds them into the market over a window you choose. The point is to avoid announcing yourself. Drop $400,000 of a mid-cap perp into the book in one click and you walk the price several levels against yourself before the fill completes. Feed the same size in over four hours and you pay something close to the average market price across that window instead.

Hyperliquid shipped TWAP in early 2024 with a fixed 30-second cadence and a 24-hour ceiling. The July 2026 update rewrote all three of the constraints that made it awkward for anything other than same-day execution.

What Changed in the July 2026 Update

Three parameters moved. Each one removes a specific reason traders used to give up on TWAP.

The old design had an awkward interaction between the fixed interval and the per-suborder minimum. A 30-minute TWAP always produced roughly 60 suborders, so a $500 order tried to send $8.33 slices and got rejected outright. Traders had to reverse-engineer a safe total from the duration, which is a strange thing to ask of an execution tool.

Now the exchange solves for the interval instead of the slice size. You give it size and time, it works out a cadence that keeps every suborder above $10.

How the Dynamic Interval Works

Hyperliquid divides your total size by the number of suborders it can fit, then spaces those suborders across the window. Thirty seconds is the tightest it will ever go. From there, the spacing widens as the ratio of duration to size grows.

The official examples make the mechanic obvious:

Same notional, wildly different execution shape. The one-hour version runs at the 30-second floor because $10,000 across 120 slices leaves plenty of room above the minimum. The four-day version cannot use 30-second slices, because 4 days at 30-second intervals is 11,520 suborders and each one would be under a dollar. So the interval opens up to about six minutes and the slice size lands on the $10 floor.

Working Out Your Own Cadence

Rough arithmetic gets you close enough to plan with:

1. Divide your total notional by $10 to get the maximum number of suborders the size can support
2. Divide your running time in seconds by 30 to get the maximum number of suborders the window can support
3. The smaller of the two is roughly what you will get, and your interval is running time divided by that number

A $2,000 TWAP over 6 hours: size supports 200 slices, the window supports 720. Size is the binding constraint, so expect around 200 suborders of $10 spaced roughly 108 seconds apart.

The $100 Minimum and Why It Matters

A $100 floor on total size is low enough that TWAP stops being an institutional-only tool. On a $100 order over a five-minute window you get ten suborders of $10 each, thirty seconds apart. That is not going to save you much on a liquid market like BTC where the book absorbs $100 without blinking.

Where it does earn its keep is on thin markets. Newly listed perps, small-cap spot pairs, and some of the equity and commodity markets on trade.xyz have books where a few hundred dollars genuinely moves the price. A retail-sized TWAP on a thin book can beat a market order by more than the fee difference.

Slippage Cap, Catch-Up, and Incomplete Fills

Each suborder carries a maximum slippage of 3%. It will not chase the price beyond that band to complete a slice.

When a suborder underfills, the TWAP falls behind its target pace and tries to make up the shortfall on later slices. That catch-up is capped at 3 times the normal suborder size, which stops a stalled TWAP from dumping a huge block the moment liquidity returns.

If liquidity stays thin for long enough, the window closes with the order partially filled. There is no automatic extension.

There is one more edge case worth knowing: TWAP suborders do not fill during the post-only period of a network upgrade. Upgrades are scheduled and announced, but a 7-day TWAP has a much higher chance of straddling one than a 2-hour TWAP ever did. Budget for a gap in execution.

Trigger Price and Max/Min Price

Two price parameters turn TWAP from a passive drip into something you can arm and leave alone. Both read the mark price, not the last trade.

Trigger price holds the TWAP dormant until the mark reaches your level, then starts the clock. Combined with a 7-day window, this is how you queue an accumulation plan for a level you expect the market to reach sometime next week without babysitting the screen.

Max price on a buy, or min price on a sell, kills the TWAP mid-execution if the mark crosses your level. Use it to abandon an entry that has run away from you, or to stop feeding sell size into a collapse.

A worked example. You want 150,000 USDC of ETH accumulated, but only below $3,200, and you want out of the plan entirely if ETH breaks $2,900. Set a buy TWAP with a $3,200 trigger, a $2,900 min price, and a 5-day running time. Nothing executes until ETH touches $3,200. From that moment the TWAP drips across the remaining window until either the size completes or ETH loses $2,900.

Randomize

An optional randomize setting varies each suborder by up to 20% either side of the standard slice. The average duration and total size stay the same, but the pattern gets harder to fingerprint. Anyone running a 7-day TWAP on a market where other participants watch the tape should have this on.

Picking a Running Time

The longer the window, the closer your average fill tracks the market average and the less impact you have. You also carry more exposure to the market simply moving while you execute. That tradeoff is the whole decision.

What Multi-Day TWAPs Add to Your Risk

The 7-day window is the headline feature, and it introduces a few things a 2-hour TWAP never had to think about.

• Funding accrues on filled size. Every hour, the portion already filled pays or receives funding. A week-long buy TWAP on a market with persistently positive funding is paying the long side rate on a growing position for days. Price the funding into the plan, not just the slippage.
• Margin moves under you. As the position builds, so does the margin requirement and the liquidation level. If you are running the TWAP near your collateral limit, a move against you mid-window can put the filled portion at risk. The liquidation guide covers the mechanics, and running isolated margin keeps the exposure ringfenced.
• Trend risk dominates slippage savings. Over seven days, market direction swamps the few basis points a TWAP saves on impact. A max/min price is not optional on long windows. It is the thing that stops a good execution plan from becoming a bad position.
• Upgrade windows. Scheduled network upgrades pause suborder fills. On a 7-day order this is more likely than not.

TWAP vs Scale Orders vs Manual Laddering

Hyperliquid gives you three ways to break up a large trade, and they solve different problems.

The clean way to think about it: a TWAP is time-certain and price-uncertain, a scale order is price-certain and fill-uncertain. If you must be in the market by Friday, use the TWAP. If you only want the position at your price and are content to miss it, ladder limits and collect the maker rebate instead.

Fees on TWAP Orders

Suborders cross the book, so they pay the taker rate: 0.045% on perps and 0.070% on spot at the base tier, before any discounts. HIP-3 builder markets charge 0.09% taker.

Two things reduce that. A referral code takes 4% off for life, and HYPE staking tiers stack on top, from 5% at Wood up to 40% at Diamond. Every suborder in a 1,000-slice TWAP gets the same discount, so the effect compounds across the whole order. Full breakdown in the fee tiers guide.

Placing a TWAP via the API

Bot operators place TWAPs through the twapOrder action rather than the standard order endpoint. The fields:

The m field is the one the update touched. It used to top out at 1440. With the 7-day ceiling it accepts values up to a full week, so check your SDK version enforces the new bound rather than the old one before you send a multi-day order and get a rejection.

Cancellation uses a separate twapCancel action referencing the TWAP id returned on placement. See the Hyperliquid API guide for authentication and signing, and the trading bot setup guide for a working harness.

Common Mistakes

• Treating TWAP as a fee saver. It is an impact tool. You still pay taker on every slice.
• Running a long TWAP with no max/min price. Over days, direction matters more than execution quality.
• Ignoring the fill report. Partial fills are normal on thin books and leave you with a position you did not plan.
• Forgetting funding on multi-day windows. The filled portion accrues from the moment it fills, not when the window ends.
• Using a TWAP to exit in a fast market. A liquidation cascade is exactly when the 3% slippage cap stops your suborders from filling. If you need out now, take the market order and the slippage.