The Trust Score: how every prop firm is scored, out of 100

#1Propr.xyz86.7/100Excellent

Still the most complete published rulebook in the category, and the only one here that publishes a dated changelog at all. What that log currently covers is narrower than the rules beside it, and we read the page ourselves on 10 September 2026 rather than trusting an older note: the header shows v1.0, the v1.0 launch entry and the v1.0.1 update (both 25 February 2026) record four account sizes, a $200,000 aggregate cap and a $50 payout minimum, and the last logged entry is v1.0.2 (18 March 2026) — while the live rules on the same page sell six sizes up to $200K, a $300,000 aggregate cap and a $20 minimum. We are re-auditing our own older references to later version numbers before drawing any conclusion from that gap. Propr is also the first firm here that actually paid us — 119.79 USDC, settled in two minutes. No consistency rule, no time limit, no bot ban, and a transparency dashboard that publishes a measured pass rate (15.2% on paid challenges, 2,366 of 15,583; 13.3% on free trials) most firms would never show. Payouts re-rated upward on 26 August 2026: the earlier markdown said its payer was an ordinary wallet with no contract code, but Propr now publishes an Address Directory whose Payout Distributor is a verified contract on Ethereum (0x6e81…7acf24, 3,864 payout transactions at our reading) — and its dashboard publishes lifetime revenue against lifetime payouts, making it one of the two firms here whose payout ratio a reader can compute, and the only one whose payouts also run through a contract you can audit transaction by transaction (Carrot Funding joined that column on 10 September 2026; its own reserve is documented and readable too, in a Safe on HyperEVM). Still marked down on two counts: a 20% price rise on the 2-Step 100K shipped with no announcement, a disclosure failure under our own rules criterion; and the distributor contract holds ~no balance — payouts are funded per transaction, so unlike Hypernova's pre-funded vault there is still no reserve you can read before it pays.

#2Hypernova85.7/100Excellent

The only firm in this table whose profit split we have verified by arithmetic rather than by reading a page: two payout receipts reconciled against its Vault contract at exactly 80%, to the cent. Its published reserve matched our own on-chain reading to the cent when we checked (8 August 2026, $770,556); a later reading, on 18 August 2026, showed $1,001,060.62 — it has fallen since, and the current dated figure is the one shown at the top of this page — after the firm announced a top-up on X and the chain confirmed it — 260,395.00 USDC landed in the reserve wallet against the “+$260,000” claimed. Account rules, equity updates and status changes are written to a public contract — 2,308 events in fifteen days — so a third party can reconstruct what happened to an account without asking anyone. Held back by three things we found ourselves: the rule-enforcing contract is an upgradeable proxy whose owner can replace the implementation and whose signer threshold is undisclosed, which makes the claim that its rules are 'immutable' too strong; its rulebook contradicts itself on bot restrictions; and a portfolio desync during our own session showed 'FLAT' and '$0.00' while a $20K position was open, which is a dangerous thing for a terminal to display. We have traded it, and between 28 and 31 August 2026 we took nine payouts from it — $3,494.69 requested on our $25K funded account, every one marked Paid with a public Arbiscan hash, receipts issued instantly. Spot-checked on-chain: the $699.43 request settled 559.546701 USDC (tx 0xf0c0…ac88) — exactly 80% to the cent, the third arithmetic reconciliation of this split against the public ledger. The last hold-back — never having been paid — is gone; what remains between this score and the top of the table sits in the rules criterion, not here.

#3Carrot Funding73.2/100Good

Bought and traded with our own money on 10 September 2026. Its four published payout hashes all settle on-chain, and its 'no fee markup' claim measured exactly 0.045% per side on two round trips — the only fee verification in this table. We also settled the question its rulebook fumbles: with two positions open, the account tracks equity apart from balance and charts equity against the drawdown floors, so floating losses do breach you — and the interface states the floors in absolute dollars rather than percentages to interpret. Against that: its own cheatsheet still says the opposite, the FAQ and the rulebook give different maximum losses and different forbidden-practice lists, leverage caps at 5x, automation, AI tools, copy trading and hedging are all banned, and the Points page states 279 traders in total with a $69 weekly pool. Two criticisms we published earlier that day were retracted the same evening after the firm answered us: its vault IS documented — a full address table with named roles, which we had not read — and the $788,521 of "funding capital" on its Analytics page is an internal risk metric, not a reserve, so there was no gap to report. Our challenge also does hold a real NFT: token #550, minted on Arbitrum at the hour we paid. The evaluation itself breached that same evening on the daily limit — two BTC/USD longs at 5x carried overnight with no stop loss, force-closed at 19:30:42, $5,000 to $4,711.22 — and that breach is ours, not the firm's; nothing here is scored against Carrot for it, and the score stays at 73.2. Both errors were ours.

#4Solana Funded69.8/100Fair

Up to 90% and Solana-native, which suits memecoin traders specifically. Documentation is average for the category and the track record is short. (Split updated 15 August 2026; Solana Funded's site now states 90% throughout.)

#5Breakout69.5/100Fair

Kraken ownership makes it the safest bet on survival by a wide margin. But settlement is off-chain on a centralised terminal, which costs it heavily on verifiability — the whole reason to pick an on-chain firm.

#6HyperPNL64.5/100Fair

We bought a 1-Step Flex $5,000 evaluation for $42 on 9 September 2026, funded it in USDC on Arbitrum and traded it the same day — so this score is no longer desk research. What the money bought works: a real Hyperliquid order book, at least 92 instruments, 5x asset leverage that let a $5,000 account carry $24,594 of notional, and limits shown as figures you can check ($4,850 daily floor, $25 profitable-day trigger). Two of our own earlier criticisms fell: the 1-Step requires no minimum profitable days at all — the counter reads 0/0, so the checkout was right to show no such line — and the firm does publish a named operator, Hyperpnl Ltd. in the Cayman Islands, behind the login. What the money did not buy is the thing the homepage sells. We traced our own $42 purchase end to end: the single on-chain event was a deposit into RelayDepository, a third-party bridging contract, and the fee itself was debited off-chain in HyperPNL's internal ledger. No Payment Router, no Payout Treasury allocation, no revenue split, no referral fee, and no published address for any of the three contracts the docs name. The one field that looks like an on-chain record, labelled Tier, is not an Arbitrum transaction. That is a disclosure failure, and we score it as one: documenting contracts you never expose costs points on rules. It is NOT scored against payouts, because we have never requested one from HyperPNL — the purchase rail is not the payout rail, and this grid only moves on what it has measured. The engine does enforce: our $25,000 demo was closed automatically on 9 September at 12:18 with the cause named, Daily Loss Limit, at -3.3%. No payout taken yet: the account is in Phase 1 and negative.

#7GT Funded18.3/100Avoid

The cautionary tale of the category. Advertised the best headline numbers (up to 92%, $300K), then went completely dark in June 2026 with no statement and nothing for its funded traders. The July relaunch on a new domain is a waitlist page whose own figures contradict each other. Scores near zero on the criteria that matter.