Axiom vs FOMO

Ranking is by raw builder fee revenue in USD, not by value for users. Bigger number means more fees collected via the on-chain builder code field. For the trader cost perspective, switch to the Effective fee bps companion tab. Native Hyperliquid UI orders carry no builder code and are excluded. Figures describe the last complete UTC day published on Hyperliquid's builder fills feed, so they run 3 to 27 hours behind the chain.

Volume is what each app's DeFiLlama dexs adapter attributes to it per UTC day, over every chain that adapter covers. Commission is the app's own cut from its fees adapter (dailyRevenue), not the total fees paid on the trade: the venue underneath takes its own share, and counting that would roughly double the figure. The take rate divides the two over the same days, so it is a floor where an adapter covers fewer chains for fees than for volume. Rows end on each app's own latest closed day.

Volume and revenue from DeFiLlama's DEX and fees APIs, 24h and 7d totals refreshed every 30 minutes. Terminal volume is a floor: a swap counts only when its transaction pays the terminal's fee wallet and Dune decodes the venue. Revenue (dailyRevenue, LP fees stripped) counts every fee reaching those wallets. Their ratio is a take rate this bench does not rank on: it moves threefold across a fortnight for one app. Bench 201 publishes it daily. Terminal and launchpad volume are not additive.

Volume is what each app's DeFiLlama adapter attributes to the app per UTC day, over every chain the adapter covers: FOMO is Solana only on DeFiLlama, GMGN spans ten chains. Each row ends on the app's latest closed day: most adapters are Dune queries that publish a day 10 to 20 hours after it closes, so before mid-day UTC some rows are dated one day behind, and GMGN's adapter skips some days. DeFiLlama can restate the last day for about a day; the series is re-read hourly.

Random sample (150 a day per row, Solana and EVM alike) of the swaps each terminal routed, every attempt counted for the fail rate. Reference = the pool's state before the swap (its reserves, or the previous trade within 60 s); swaps without one keep their split, no loss figure. Published from 50 priced swaps pooled over the product's chains at 25 or more (a single-chain product from 25), ranked from 100; the median's 95 % interval is in the JSON. Compare medians, not single swaps.

Same sample, window and reference as bench 268; read that page for how a swap is priced. Here each swap's loss has the terminal's fee subtracted before the median is taken. Subtracting the two published medians gives a different number wherever an app's fee varies across its own swaps, which is why this is measured, not derived. A swap reads negative when a fee was rebated or the reference drifted the user's way; those stay in. Compare medians with their intervals, never single swaps.