A record-setting session for cross-exchange arbitrage — 417 flagged spreads including a blistering 40.80% gap on H between Bybit and OKX, plus double-digit windows on BEAT, BASED, LAB, and MOODENG. Full breakdown of buy/sell legs, fee-adjusted profit math, and exchange-pair patterns for August 22, 2026.
417 flagged spreads crossed the desk today, and the top of the board is loud. H printed the session's headline number — a 40.80% gap between Bybit ($0.062239) and OKX ($0.064490) — followed by BEAT at 39.50% between Bybit and Gate Futures, and BASED at 39.46% between Bybit Spot and OKX Spot. When three separate assets clear the 39%+ threshold in the same session, that's not noise from a single illiquid pair misfiring — that's a market-wide liquidity fragmentation event, and it's exactly the kind of session arb desks wait for.
The shape of today's board is worth noting before diving into individual legs: this isn't a spot-vs-spot story alone. Futures venues (Gate Futures, Binance Futures) are showing up repeatedly on the sell side against spot buy legs on Bybit, OKX, and KuCoin — meaning a chunk of today's spread is basis-driven rather than pure exchange-to-exchange mispricing. That distinction matters enormously for execution: a spot/futures basis spread behaves differently under funding-rate pressure than a pure spot/spot dislocation, and treating them the same way is how arbers give back gains they thought they'd locked in.
With volume and pressure metrics reading flat at $0.0M across the board today, this report leans on the price legs and spread percentages as the primary signal — treat the absence of volume confirmation as a caution flag on every trade below, not as a reason to skip the analysis. Let's get into the top five.
Bybit is the dominant buy-side venue across today's top opportunities — it appears as the buy leg in H, BEAT, and BASED, and as the sell leg in COTI. That's a consistent signal: Bybit's spot pricing on mid-and-lower-cap alts is running persistently cheap relative to OKX and Gate today, which is either a temporary liquidity/inflow imbalance specific to Bybit or a sign that arb bots haven't yet fully closed the gap on these particular tickers.
OKX shows up almost exclusively as the sell-side venue (H, BASED, USELESS), reinforcing the pattern that OKX pricing is running rich relative to Bybit right now. The Bybit-buy / OKX-sell pairing alone accounts for two of today's top three spreads — if this pattern holds into tomorrow's session, that specific pair deserves a standing watch rather than a one-off check.
Futures venues (Gate Futures, Binance Futures) consistently sit on the sell side against spot buy legs (BEAT, LAB, ORDI, RED all follow this shape), which as noted above means basis risk is baked into roughly a third of today's opportunity list. KuCoin appears twice as a buy-side venue (LAB, MOODENG) — both against smaller or futures-side counterparties — consistent with KuCoin running a persistent, if modest, discount on smaller-cap names versus tier-1 venues. Bitget shows up on both sides (buy in COTI, sell in BROCCOLI and USELESS), suggesting no consistent directional bias there — treat Bitget spreads case-by-case rather than assuming a pattern.
The spreads on this board split cleanly into two trade types, and conflating them is the fastest way to under-perform. The spot/spot pairs (BASED, USELESS, COTI, BROCCOLI) are fast-decay opportunities — book imbalances that arb bots and manual traders alike start closing within seconds to low minutes. These reward speed and pre-funded capital far more than size; trying to route a large position through a slow manual transfer defeats the purpose, since the spread will likely have compressed by the time funds land.
The spot/futures basis pairs (BEAT, LAB, ORDI, RED) decay more slowly because part of the spread reflects funding-rate-driven basis rather than pure order-book imbalance — that basis can persist for hours, which is why cash-and-carry books can run size into these rather than racing the clock. But size cuts both ways here: larger notional on the futures leg means more slippage on entry and exit, and funding can flip against the position if held too long.
Position sizing recommendation: for the fast-decay spot/spot pairs, size to what the thinner side of the book (typically the sell venue) can absorb without moving the realized price by more than 15-20% of the quoted spread — for BASED and USELESS specifically, that likely caps clean execution in the low-to-mid five figures given today's venues. For the basis trades, size to your comfort with holding overnight funding exposure, not to the headline spread percentage — the 39.50% on BEAT is not a number you capture in one clean transaction.
Walking through BASED as the cleanest example (spot/spot, both legs liquid, no futures basis complication): buy at $0.083170 on Bybit Spot, sell at $0.086630 on OKX Spot, gross spread 39.46%. On a $10,000 position: buy-side taker fee at a typical 0.10% costs $10; the resulting BASED is sold on OKX Spot, and at 0.10% taker fee on the ~$13,946 sell-side notional (10,000 grown by the spread), that's roughly $13.95 — round-trip trading fees land near $24. Withdrawal fee to move BASED from Bybit to OKX depends on the network it settles on, but assume a mid-range $3-8 flat fee for a standard token transfer. Net: roughly $3,946 gross profit minus ~$24 in trading fees minus ~$5 in withdrawal costs, landing net profit around $3,917, or roughly 39.17% net on the $10,000 base — fees barely dent a spread this size.
That said, spreads this size are the exception, not the baseline — most sessions produce spreads in the 0.3%-2% range, and that's where fee math actually determines whether a trade is worth taking. On a typical 1% spread with the same $10,000 position: gross profit $100, minus ~$20 in round-trip taker fees (0.10-0.10%), minus a withdrawal fee that can run anywhere from $1 (stablecoins on cheap networks) to $10+ (tokens on congested chains) — net profit lands somewhere between $70-$79, or 0.7-0.79% net. Push the withdrawal fee higher, or use maker-side limit orders where taker fees run closer to 0.02-0.05%, and the math improves meaningfully.
Minimum spread worth chasing: as a rule of thumb, don't route capital into anything under roughly 0.5% gross spread unless fees on both venues are confirmed sub-0.05% (maker rebates, VIP tiers) — below that threshold, withdrawal fees and any price slippage during the transfer window routinely erase the entire edge. Today's top ten are all well clear of that floor, which is what makes this a genuine arb-hunting session rather than a fee-farming trap.
Withdrawal delays are the single biggest execution risk across every opportunity above — a 30-40% quoted spread means nothing if the transfer between venues takes long enough for the gap to close, and smaller/newer venues (Bitunix, Toobit-tier exchanges) are historically the slowest and least predictable on this front. Confirm current withdrawal processing times on the specific venue before committing capital, not from memory of how that exchange behaved last month.
Low liquidity on the sell side is the second major risk, particularly for LAB (KuCoin buy leg) and MOODENG (Bitunix sell leg) — quoted top-of-book prices on thinner venues are frequently not achievable for meaningful size, and market-order execution can slip 5-10% below the quoted price on a book this thin, which directly cannibalizes the arb margin.
Basis-trade funding risk applies to BEAT, LAB, ORDI, and RED — any position left open on the futures leg overnight is exposed to funding rate payments that can silently erode or reverse the captured spread. And a broader structural flag: today's volume and pressure metrics all read $0.0M, meaning none of these spreads come with confirmed depth-of-market backing in the data feed. Treat every quoted price above as a snapshot, not a guaranteed fill, and size conservatively until liquidity confirmation improves.
Watch the Bybit/OKX pair first thing tomorrow — with H, BASED, and (on the sell side) USELESS all showing OKX running rich against Bybit today, that imbalance may not fully close overnight, especially on lower-liquidity hours in the Asia session when book depth on both venues thins out. The Bybit/Gate Futures and KuCoin/Binance Futures pairs are worth a standing watch too, given the basis-driven spreads on BEAT and LAB — check funding rates on both futures venues at the next funding interval, since a funding flip could either compress or re-open those gaps fast.
Best times to watch: exchange handoff windows (roughly 00:00 and 08:00 UTC, where Asia/EU and EU/US liquidity providers rotate) are historically when book-depth gaps like today's H spread tend to open, since market-maker coverage briefly thins. Keep an eye on smaller-cap tickers with recent listing or delisting news on any single venue — that's the most common trigger for a 30%+ spread like the ones topping today's board, and it tends to cluster around news events rather than appearing randomly.
417 spreads, a 40.80% headline number, and a board that rewards fast hands on the spot pairs and patient basis books on the futures pairs. Fund your accounts, check the withdrawal queues before you need them, and don't let a wide quote talk you into a fill you can't actually get. Arbitrage Hunter — August 22, 2026.