# The Leverage Trading Platforms List, Sixteen Venues Deep The Leverage Ladder desk · Updated: 24 September 2026 · Tested: 22 September 2026 ## The one line This leverage trading platforms list covers sixteen venues read on 18 September 2026: six that publish a notional band, four that print a cap without one, six with no resting book. EVEDEX heads the ranked rows, printing 200x on BTC to $50,000 of notional. At that rung, half a percent against a position ends it. Weights: Depth under the position 35 · Cost on the levered notional 25 · Where the fill is recorded 20 · What the record shows 15 · How far the ladder reaches 5. One formula for every plate on this site, fixed on 18 September 2026: https://aurascopeapp.com/method ## The leverage trading platforms list, the six ranked venues by documented cap Six venues, sorted by the documented cap and then by the notional that cap holds to: coverage, not a verdict. A row exists here only where the venue publishes both figures — the multiple and the position size it survives — because a cap with no step beneath it cannot be set against another venue's. Four venues that print one without the other are [named below](#four-caps-with-nothing-underneath): Lighter, Paradex, dYdX and Aevo. Six more leave no resting book for the depth criterion to read, and are [further down again](#listed-and-not-scored). No total sits beside a row here; the six totals are written out on [the exchange plate](/best-leverage-crypto-exchange) and tabled criterion by criterion on [the method plate](/method). | # | Venue | Leverage cap | Maker / taker | Perpetual markets | Model | |---|---|---|---|---|---| | 1 | EVEDEX | 200x to $50,000 | 0.015% / 0.045% | 52 | Hybrid | | 2 | Aster | 200x to 400 USDT | 0% / 0.040% | 578 | Own layer 1, operator-run | | 3 | edgeX | 100x to $1 million | 0.040% / 0.045% | 172 | Hybrid | | 4 | ApeX Omni | 100x to $100,000 | 0.020% / 0.050% | 125 | Hybrid | | 5 | Extended | 50x to $4 million | 0% / 0.025% | 325 | Hybrid | | 6 | Hyperliquid | 40x to $150 million | 0.015% / 0.045% | 324 | On-chain order book | Venue terms read 18 September 2026; entry-tier fees on the BTC perpetual. The leverage column pairs the headline cap with the notional it holds to. No total is printed on this plate: a ladder is not one of the five things the formula weighs, and a column of totals falling while the row numbers climb asks a reader to believe two sequences at once. Every venue's total, with each criterion standing beside it, is tabled on the method plate: https://aurascopeapp.com/method ## Quoted, with dates > "The max allowed leverage of a pool decreases as the total open interest of the pool increases." — GMX documentation, 18 September 2026. https://docs.gmx.io/docs/trading/order-types/ > "The maximum leverage available when opening a position is 250x for SOL, ETH, and wBTC." — Jupiter Perps documentation, 18 September 2026. https://docs.jup.ag/user-docs/trade/perps/faq > "Degen pairs are high-leverage variants offering a fixed 500x leverage." — gTrade documentation, 18 September 2026. https://docs.gains.trade/gtrade-leveraged-trading/fees-and-spread ## 01 · EVEDEX EVEDEX tiers BTC by size: 200x to $50,000 of notional, 150x to $75,000, 100x to $200,000, then 75x, 50x and 25x as far as $716,000. ETH and SOL share that table; XRP, gold, silver and crude oil sit at 100x and the rest of the list at 75x or 50x (EVEDEX trading terms, checked 21 September 2026). - For: 0.015% maker / 0.045% taker, on a median book of $18.9 million. - Against: 52 perpetual contracts, the shortest list of the six ranked here; Positions reach Arbitrum in batches, no operator-free withdrawal documented; The BTC table ends at $716,000 of notional and prints nothing above it. - Not for: a position past $716,000 of notional, where the table ends. ## 02 · [Aster](https://www.asterdex.com) The Aster Pro book holds 200x for the first 400 USDT of notional, then 150x to $300,000 and 100x to $800,000 ([Aster trading rules](https://www.asterdex.com/bapi/futures/v1/public/future/simple/symbols), checked 18 September 2026). - For: 578 markets and a median BTC spread of 0.01 bps. - Against: The 200x bracket covers 400 USDT of notional; Closed chain code and no external validators in phase one. - Not for: anyone sizing from a headline. ## 03 · [edgeX](https://pro.edgex.exchange) edgeX writes BTC as risk tiers: 100x under $1 million of position value, 75x under $3 million, 50x under $5 million ([edgeX metadata](https://edgex-prod-v2.edgex.exchange/api/v2/public/meta/getMetaData), checked 18 September 2026). - For: The deepest book we measured, a median $29.2 million, and an exit past the operator. - Against: 0.040% maker fee, where two venues in these rows charge makers nothing; No bug bounty is published. - Not for: makers at the entry tier. ## 04 · [ApeX Omni](https://omni.apex.exchange/) ApeX Omni allows 100x on BTC inside $100,000 of position value, adding a point of initial margin per further $100,000 ([ApeX Omni docs](https://apex-pro.gitbook.io/apex-pro/apex-omni/perpetual-contracts-open-interest-and-leverage), checked 18 September 2026). - For: A forced withdrawal to layer 1 out of the zkLink contracts. - Against: Median BTC depth of $1.8 million, the thinnest book above 50x here; 0.050% taker fee, charged twice on the levered notional. - Not for: cost-sensitive takers. ## 05 · [Extended](https://extended.exchange/) The Extended margin schedule gives BTC 50x to $4 million of position value and 25x to $8 million ([Extended margin schedule](https://docs.extended.exchange/extended-resources/trading/crypto-markets/margin-schedule), checked 18 September 2026), settled on Starknet. - For: 0% maker / 0.025% taker, the cheapest paid round trip here, and a $500,000 bounty. - Against: 278 of its 325 markets are quoted on request, with no public book; The advertised 100x is a currency pair, not BTC. - Not for: anyone who needs a public book. ## 06 · [Hyperliquid](https://hyperliquid.xyz) Hyperliquid prints a modest headline figure and the longest reach: 40x on BTC to $150 million of notional ([Hyperliquid margin table](https://api.hyperliquid.xyz/info), checked 18 September 2026). - For: Open interest of $8.7 billion on 22 September 2026, the largest here. - Against: No operator-free withdrawal documented, searched 18 September 2026; Its own vault lost about $4.9 million in November 2025 and $4 million in March 2025. - Not for: a small account chasing a multiple. ## Four caps with nothing underneath These four run an order book, publish a BTC cap and stop there. What none of them prints is a step where that multiple gives way to a lower one, and the size it gives way at is the figure every row in the table above is sorted on, so ranking them beside a venue that does publish it would mean inventing the number the venue itself left out. - **Lighter**: a two percent initial margin on BTC, which reads as 50x, with no notional tiers in the contract specifications. Standard accounts trade at no fee behind 300 ms of added taker latency, and the venue was down for four and a half hours in October 2025. - **Paradex**: the same two percent, bounded by a position limit of 300 BTC rather than by a schedule of steps. Retail orders pay nothing behind a 300 ms speed bump, and corrupted state in January 2026 set off liquidations that ended in a $650,000 refund. - **dYdX**: two percent again, with no tier published behind it. The bounty runs to $1,000,000 for a critical report, and the October 2025 chain halt left stale prices and $462,097.79 of trader losses in a compensation proposal. - **Aevo**: a five percent initial margin, 20x, inside a $10 million position cap. Five audits stand behind it, and a December 2025 exploit of the team's legacy vaults cost about $2.7 million. A reader who wants one of these four is not being warned off it. The ladder is simply not a thing these four publish, and this site is a ladder. ## Listed and not scored Four of these six price from an oracle against a pool or a vault, one shares a single book with every front-end built on it, and one is paused, so the depth criterion has no resting orders to read and no total is set. Each publishes a cap, and each documents drawbacks to weigh against it. - [**gTrade**](https://gains.trade): a fixed 500x on its degen pairs, 178 markets, settled against shared vaults. A liquidation sends the whole collateral to the vault, and a stock position cannot be opened, closed or edited while the market behind it is shut. - [**Jupiter Perps**](https://jup.ag/perps): 250x on three markets, filled from one pool. Borrow fees pull the liquidation price toward the market with no price move at all, and limit orders cannot be placed once utilisation of a market passes 80%. - [**GMX**](https://gmx.io): 100x on 103 markets, filled by keepers against pools at an oracle price. The cap falls as open interest against a pool grows, and auto-deleveraging closes profitable positions in part or in full to keep that pool solvent. - [**Ostium**](https://www.ostium.com): 100x in live market data against 200x in the documentation, across 83 instruments. A market order is refused while a non-crypto market is closed, and a liquidation keeps whatever collateral is left. - [**Orderly**](https://orderly.network): 100x on one book shared by every front-end built on it. Each front-end sets its own fee above the base rate, so the venue has no single price, and the operator reads the full content of an order before it is matched. - [**Drift**](https://www.drift.trade): 20x on the live market against the 101x its own page still advertises. The programme has been paused since the April 2026 exploit that cost $295 million, and no balance carried over to the fork that replaced it. ## What was read, and what was measured Five criteria fixed on 18 September 2026, before a venue was read against them, applied to the 6 venues with a resting book ([the formula](/method)). Ladders, fees, market counts, audits and bounty pages come from each venue's documentation, read on 18 September 2026; open interest is as CoinMarketCap published it on 22 September 2026. Depth and spread are ours. Books were read on BTC, so a thinner pair behaves worse than these medians. Entry-tier fees ignore discounts, rebates and cashback. Reported open interest is self-published and is no evidence of liquidity. The six venues with no resting book were read, not measured. Placement here is paid for; the order follows the published formula and ladder. ## FAQ ### Which platforms offer leverage trading on crypto? Sixteen are listed here. Six are ranked: EVEDEX, Aster, edgeX, ApeX Omni, Extended and Hyperliquid. Four run a book but publish no notional band: Lighter, Paradex, dYdX and Aevo. Six carry no book at all: gTrade, Jupiter Perps, GMX and Ostium price from an oracle, Orderly is shared, Drift is paused. ### Order book or oracle pool, what is the difference? An order book matches one trader against another, so the exit depends on the size resting there. An oracle venue fills against a pool at a reference price, so size is bounded by pool capacity and the cost arrives as borrow and impact fees. ### How many perpetual markets do these platforms list? From three to 578. Jupiter Perps quotes SOL, ETH and wBTC only, while Aster lists 578, Extended 325 and Hyperliquid 324. The six ranked venues carry 1,576 perpetual markets between them, counted from each venue's own interface on 18 September 2026. ### Can you trade gold or shares with leverage on a crypto platform? Fourteen of the sixteen list something beyond crypto. Ostium spans six asset classes across 83 instruments, gTrade covers currencies, commodities, shares and indices, and EVEDEX carries five non-crypto classes among its 52 contracts, gold among them through a Tether Gold perpetual. ### Why do some platforms advertise higher leverage than they actually allow? Because the headline is the top of a ladder and the ladder sits in another document. Extended advertises 100x and gives BTC 50x; the Drift page still says 101x where the live market caps BTC at 20x; Ostium documents 200x against 100x live. ### What happens to my position if a protocol is paused? It stops being manageable until the pause lifts. The Drift programme has been paused since April 2026 and balances did not carry over to the fork that replaced it. The forced withdrawal at ApeX Omni and the exit past the operator at edgeX answer the same question. ### Which perps platform has the most open interest? Hyperliquid, by a wide margin, at $8.7 billion of open interest and $13.4 billion of 24-hour volume as published by CoinMarketCap on 22 September 2026. Aster follows at $1.4 billion and EVEDEX at $704.9 million, each figure reported by the venue itself. ### Is leverage trading available on Solana? Two of the sixteen settle there. Jupiter Perps runs SOL, ETH and wBTC against a pool at up to 250x, and Velocity, the fork that replaced the paused Drift programme, lists four markets with BTC capped at 20x. The rest settle elsewhere. ### Do any of these platforms run a bug bounty? Three of the six ranked do: Extended up to $500,000, Hyperliquid up to one million USDC and Aster up to $200,000. EVEDEX, edgeX and ApeX Omni had none listed on 18 September 2026. Among the four with no published band, dYdX runs one to $1,000,000 and Paradex to 500,000 USDC. ### How often do these leverage ladders change? Often enough that every figure here carries the date it was read. Venues cut caps in volatile hours and restore them unannounced, and live market data can contradict a documentation page on the same day, as it does at Ostium and at Drift. ## About this plate One published formula, fixed before any venue was scored; every figure carries the date it was read. Venue terms and leverage ladders come from each venue's own documentation; order-book figures are measured here. Corrections: corrections@aurascopeapp.com. The Leverage Ladder desk, 24 September 2026 A plate is paid for.