Top Prop Firms on Hyperliquid: Side-by-Side Comparison

PROPR, Hypernova, Iceberg and Ferm differ in their evaluation requirements, trading tools and payout rules. PROPR and Ferm provide trading APIs, Hypernova offers native Scale order controls, Iceberg provides account access without an evaluation, and Ferm lets you buy extra drawdown on an evaluation that is close to its limit.

This comparison combines their published rules with account results and observations from my use of all four platforms, plus API tests on a free Ferm Demo account. It is not an exhaustive ranking of every prop firm in the ecosystem.

Disclosure: PROPR and Hypernova each gave me one free account via my personal X account at @realJoJo. However, I also bought one PROPR account, four Hypernova accounts and my Iceberg account myself. For Ferm, I also bought one $25,000 Pro evaluation for $208, its list price at the 80% profit split, and I tested its API on a free Demo account. Ferm says its crypto prices come from Hyperliquid, relayed by Hydromancer, which publishes this post.

Account results and published rules reviewed as of 24 September 2026 (Ferm: 21 October 2026). Prices and account terms can change

What are these prop firms actually offering?

The basic offer is access to a trading account for a fee, with a share of eligible profits paid under the firm’s rules. PROPR and Hypernova offer evaluation-based routes: reach a target without breaching the risk limits, then qualify for the funded stage. Iceberg’s Instant account skips that evaluation.

“Funded” does not mean you own the advertised account balance or place every trade directly on Hyperliquid. All four describe simulated user trading in their terms or documentation. PROPR may copy trader signals into its own live positions; Hypernova, Iceberg and Ferm describe simulated execution using Hyperliquid market data. Real payouts and simulated trading can coexist.

That distinction matters. A $5,000 account is not $5,000 you can withdraw. The relevant questions are how much you pay, how much loss the rules allow, and what profit you can actually claim.

What do prop firms bring to Hyperliquid?

Prop firms give traders another way to engage with Hyperliquid’s markets through evaluations, trading tools and performance-based payouts. For the ecosystem, they can introduce new audiences to its markets and create demand for market data, analytics and trading infrastructure.

The direct trading benefit depends on how each firm operates. If a firm copies trader signals or hedges exposure on Hyperliquid, those live trades contribute to exchange activity. Simulated trades alone do not. The broader value is therefore not just potential trading volume, but more products and services built around Hyperliquid’s markets.

A quick comparison

These are published $5,000 plan examples, not a comparison of each firm’s cheapest advertised offer. The PROPR and Hypernova examples have matching headline targets and loss percentages; Ferm’s Pro plan has the same loss limits with an 11% target and a lower fee; Iceberg charges for immediate access without an evaluation.

All four examples use an 80% trader profit share, subject to their payout conditions; Ferm’s 90% option costs $67 on the $5,000 Pro plan. The fees shown are the published prices at the time of review, not necessarily what I paid. Check the selected plan’s terms and final checkout price before purchasing.

The percentages also aren’t the whole rule. Open losses, fees and funding can affect equity and trigger a breach before you close a trade.

1. PROPR: trading API and performance analytics

PROPR offers several evaluation formats, including Classic, Turbo and Pro one-step plans, plus a two-step option. It describes itself as built by XBorg and backed by SwissBorg.

Account results and trading tools

I tried the $5,000 Starter and $10,000 Explorer Classic one-step accounts. PROPR supports automated trading through its API. I connected Hermes Agent and instructed it to trade aggressively, but the strategy was unprofitable.

The analytics dashboard displayed performance by asset and day, long-versus-short results, trade duration, fees and slippage.

The account results here cover Classic one-step plans; I did not test the two-step evaluation.

Payout conditions and restrictions

In my experience, payouts took roughly six to twelve hours, and I couldn’t keep trading that account while the withdrawal was processing. That is my observed experience, not a universal processing time. The public rulebook says payouts are processed within 24 hours, with a $20 minimum, closed positions and a full-profit sweep rather than partial withdrawals.

KYC is required before funded activation and payouts. The documented API supports automated trading, but that access does not remove strategy restrictions.

There is also a documentation issue worth flagging: PROPR’s homepage allows copying between your own accounts, while its evaluation terms prohibit mirrored trades across accounts you control. Its funded terms warn that identical or off-the-shelf bots can be flagged as coordinated trading. These conflicting statements leave the permitted scope of multi-account automation unclear.

2. Hypernova: native order tools and on-demand payouts

Hypernova offers one-step assessments with different risk tiers. Its Conservative plan has a 10% target, 3% daily loss limit and 6% static drawdown; Standard increases the daily and total limits to 4% and 7% respectively, at a higher fee.

Account results and trading tools

I opened five Hypernova accounts: one provided free and four I purchased.

One account paid out more than $1,000 before it eventually breached. Seven payments marked “Paid” in that account’s payout history, dated 9–10 September, totalled $1,201.41. My other four accounts also breached. This payout subtotal should not be read as my net profit across all five accounts.

The terminal showed Scale controls and a TWAP tab. Its documentation explains scale orders as a ladder of limit orders across a price range.

Payout conditions and restrictions

Hypernova’s linked API specification explicitly says it is currently UI-only and does not expose a public API.

The payout documentation says there is no minimum and no cooldown, but the account must be eligible and flat, with no open positions or orders. It describes approximately five-second end-to-end settlement—not a guarantee for every withdrawal.

The dedicated payout documentation also lists a $10,000-of-profits daily cap per user during closed alpha, and requires another open-and-close trading cycle before repeat settlement. Its terms also reserve the right to request identity verification.

Its rulebook also prohibits copying third-party trade ideas and switching to a materially different strategy after passing the assessment. These apply in addition to the account’s loss limits.

Hypernova announced that KYC will become mandatory to receive a funded account from 30 September 2026. In-app verification opens on 25 September, through Sumsub.

3. Iceberg: no evaluation and a per-request payout ceiling

Iceberg’s clearest difference is its Instant model: pay for access and start without passing an evaluation. Its currently available sizes are $5,000, $10,000 and $25,000, priced at $167, $337 and $767. The larger advertised tiers are marked coming soon.

That makes the entry fee higher than the evaluation examples above. You are paying to skip the assessment, not buying ownership of live trading capital.

Account results and pending payout

I bought the $5,000 account for $167. At the time of writing, it showed $349.47 in unclaimed profit.

The payout screen applied the following limits and split:

My share of the payout is $200, covering the $167 account fee and leaving $33 above the initial cost. I had therefore reached breakeven on an earned-profit basis, with the payout still awaiting receipt.

Payout conditions and account lifetime

Iceberg’s documentation confirms a 5% maximum gross payout per request, measured against initial account size, before the 80/20 split. The minimum eligible profit is 1%. Claims take all eligible realized profit; partial payouts are not allowed. Any amount above the cap is permanently forfeited when the claim settles successfully.

For my $5,000 account, that explains the $250 gross ceiling and $200 trader share. Profit above the cap does not carry into the next withdrawal once the claim settles.

Timing matters too. Current documentation says the first request opens five days after the first executed trade, and subsequent requests open five days after the previous payout is received.

Accounts also expire 90 days after activation. And becoming eligible to request a payout does not guarantee immediate payment: Iceberg says requests can remain queued until its payout pool has available liquidity.

Also, Iceberg is the only platform that does not impose KYC as of now, a big win for those who values more privacy.

4. Ferm: an agent-friendly API and a drawdown top-up

Ferm offers one-step evaluations on three paid risk tracks, Beginner, Pro and Degen, plus a free Demo track that does not pay out. The $5,000 Pro plan costs $54 at the default 80% profit split, or $67 for 90%, with an 11% target, a 3% daily loss limit and a 6% static drawdown. There is no time limit and no minimum number of trading days. Ferm says every crypto fill is priced from a live Hyperliquid order book, with prices “relayed by Hydromancer”, and that fills are simulated and nothing is sent to an exchange. Unlike PROPR, it does not describe copying trader signals into live positions.

Three things stood out: an API built for bots and AI agents, an optional drawdown top-up that I did not find at the other three firms, and a PnL card that I think is the best presentation of a result among the four.

Account results and trading tools

I used two Ferm accounts: a paid $25,000 Pro evaluation, which cost $208 (Ferm’s list price at the 80% profit split) and has since breached, and a free $5,000 Demo account for testing the API. I have not requested a Ferm payout, so I cannot speak to payouts from experience. The API notes below come from Ferm’s documentation and from running the API on the Demo account, with an AI coding agent (Claude Code) placing the test orders. The rules come from Ferm’s rulebook, version 2.32 of 9 October 2026.

Ferm treats bots as ordinary users. It documents a REST API with scoped keys (account:read and order:write) and an event feed for recovering after a restart. The rate limits are generous for a polling bot: 600 requests per minute per key for reads and another 600 for orders, with standard RateLimit headers on every response so a bot can pace itself.

Its Help Center has a guide to setting up a bot, which says a bot can be a simple script, a rules engine, a market-making system or an AI trading agent. The same daily loss, max drawdown and funded rules apply to bots, and news trading is allowed.

Through the API I placed market, limit and chase orders, set stop-loss and take-profit by price, moved them on an open position, closed positions partly and fully, opened long and short positions and sent several orders in parallel; a burst of 40 order requests in 42 seconds went through without any issues.

Ferm’s PnL card is, I think, the best presentation of a trade result on any of these platforms.

After I closed a 2x HYPE long on the paid evaluation, it produced an 11-second animated card styled like a split-flap departure board. It opens with the market, side, leverage and entry price, then counts up the realized P&L while a grid draws the path of the trade, marking the low (−$48.39 on mine) and the comeback.

It peaked at +$112.90 and ended on the result: +$106.88, +1.77%, held 6 hours 35 minutes. A second pass shows what happened after I closed, with an “if you held” counter that finished at −$12.38 and the label “sold the top”.

The last frame carries the entry and exit prices (79.2622 and 80.0025) and a link to Ferm on web, iOS and Android, so the card works as a shareable result on its own.

Some limits showed up in testing:

• Leverage is fixed per instrument and cannot be changed, in the API or the web terminal: 4x on BTC, ETH and SOL, 2x on other crypto, higher on forex and CFDs. Size is the only lever.
• Calls took about one to two seconds each, and the WebSocket stream scope is listed as coming soon, so the API suits slower strategies.
• Setting take-profit and stop-loss by percentage, as in the docs’ own quickstart example, left the position with no stop and no take-profit in my test; setting them by price worked. Five orders also returned a 502 engine error within a few minutes, and the same orders succeeded when retried hours later.

The drawdown top-up

On an active evaluation, when equity gets within 2% of the starting balance above the max drawdown floor (1% on Degen), Ferm offers a one-time purchase that lowers the floor. Beginner and Pro accounts can add 1% or 2% of the starting balance; Degen accounts can add 1%. The price is 12% of the drawdown you add.

On the $5,000 Pro plan, the 6% static floor sits at $4,700, so the offer appears once equity is at or below $4,800. By Ferm’s formula, adding 1% costs $6 and moves the floor to $4,650, and adding 2% costs $12 and moves it to $4,600.

The limits matter. It is a one-time purchase per account, only on evaluations, not on funded or live accounts, and it is final once applied. The daily loss floor does not move, so a loss that touches it still breaches the account. Ferm calls the offer optional and says it is not a recommendation to keep trading. Its help article is dated 16 October 2026, so the feature is new.

The other three firms work differently. PROPR’s and Hypernova’s rulebooks say a breached account cannot be recovered, reset or appealed, and Iceberg’s documentation says reaching a loss limit closes the account permanently.

Payout conditions and restrictions

Funded accounts keep 80% of profit by default, or 90% with the checkout upgrade. Under new standard funded agreements, each payout period needs three qualifying trading days. A qualifying day means closed profit of at least 0.4% of the starting size, which is $20 on a $5,000 account. The minimum withdrawal request is $57 on a $5,000 account, because the trader’s share has a $45 floor after the split.

Each approved request can release up to 4%, then 5%, then 6% of the starting funded size, which is $200, $250 and $300 gross on $5,000. Remaining profit stays on the account for a later request. A request needs all positions closed and pending orders cancelled, and trading is locked while it is under review. Payouts settle by bank ACH or wire, or by USDC on Ethereum, Base or Solana. I did not test payout timing, and the rulebook gives no processing time for payouts.

KYC is not required to buy or pass an evaluation. It is required before the funded agreement is issued and remains a condition of every payout. Ferm asks for a government ID, a selfie and a typed residential address, with no proof-of-address document.

News trading and weekend holding of crypto are allowed. Ferm prohibits latency arbitrage, tick sniping, news straddle abuse, cross-account hedging and wash trading. Its fraud engine flags accounts for human review before any action is taken. Funded accounts add a 1.2% live-loss cap per trade idea.

The main differences

• PROPR provides a documented trading API and offers one- and two-step evaluations. Its payout rules specify a $20 minimum and processing within 24 hours.
• Hypernova provides native order tools but currently no public API. Its payout documentation states no minimum or cooldown, subject to eligibility and the closed-alpha daily cap.
• Iceberg removes the evaluation stage. In exchange, its Instant plans have a higher entry fee than the evaluation examples above, a 5% gross payout ceiling per request, five-day eligibility periods and a 90-day account lifetime.
• Ferm provides a documented REST API and says bots and AI agents are welcome. It is the only one of the four whose published rules let you buy extra drawdown on a live evaluation, once, and its payouts need three qualifying trading days per period, with a $57 minimum request on $5,000.

None of this makes my results a prediction of yours. I breached accounts on PROPR, Hypernova and Ferm. A few successful payouts do not establish long-term profitability, and advertised funding size is not a substitute for reading the loss limits.

For broader context—not a prop-firm success-rate comparison—our research on Hyperliquid frontend traders examines realised trading outcomes and why trader selection matters.

Account size alone does not determine what a trader can withdraw. Access fees, loss limits, profit splits and payout conditions determine how these plans differ.

Sources

#1: Propr Model

#2: Propr Rules

#3: Hypernova Rules

#4: Hypernova Payouts

#5: Iceberg Model

#6: Iceberg Payout Requests

#7: Ferm Rulebook

#8: Ferm Max Drawdown Add-on

#9: Ferm Agent API

#10: Ferm Trading Bot Guide

#11: Ferm Transparency

#12: Ferm Plans and Prices