September 12 arb scan surfaces 65 cross-exchange spreads, led by a 16.67% 龙虾 gap between Gate Futures and Exchange24. Zero reported volume across the board is the real headline — every window here needs a manual depth check before size goes on.
Sixty-five arbitrage windows crossed the desk today, and the top of the board belongs to 龙虾 — a 16.67% spread between Gate Futures ($0.114512) and Exchange24 ($0.124664). That's not a typo-tier discrepancy you see on major pairs; it's the kind of number that only shows up on thin, low-cap tickers where one venue's order book hasn't caught up to another's. Three of today's top ten prints belong to 龙虾 alone, which tells you this wasn't a one-off blip — Exchange24 was running persistently rich against every venue it touched with this asset all session.
The rest of the board reads like a standard mid-cap dislocation day: BEAT and LAB both posted spreads near 12% against the OKX/Binance Futures axis, LSK and Q showed the Bybit/Binance Futures pairing twice, and the tail of the list settles into the 6-7% range where most of the real, executable volume tends to live. One thing stands out hard in the totals block: pump volume, dump volume, buy pressure, and sell pressure all read $0.0M. That's either a data-feed gap on the volume side or these opportunities are genuinely thin-air quotes with nothing behind them — either way, it changes how this report should be read. Treat every spread below as a signal to go verify order-book depth manually, not as a green light to route size.
Buy side: Gate Futures at $0.114512. Sell side: Exchange24 at $0.124664. This is the widest print of the day by a wide margin, and it's a repeat offender — 龙虾/Exchange24 shows up three separate times in today's top ten, which suggests Exchange24's book for this ticker was structurally detached from the rest of the market rather than flickering on a single stale tick. With zero reported volume in the feed, though, there's no way to confirm how deep either book actually is at those prices. My take: this spread is real on paper and probably real on a screen, but at sub-$0.12 price levels on what reads as an obscure or newly-listed token, the realistic executable size before slippage eats the edge is likely small — think hundreds, not thousands, of dollars per clip unless someone confirms deeper resting liquidity first. Worth chasing for a test-size fill, not worth backing the truck up on blind.
Buy side: OKX at $0.087100. Sell side: Binance Futures at $0.097500. This is the cleanest pairing on the board — both venues are top-tier, deeply liquid exchanges, which meaningfully de-risks the execution side versus the exotic-venue plays elsewhere on this list. The concern here isn't counterparty or book depth, it's the spot-to-futures nature of the trade: selling into Binance Futures means you're either running a basis trade (long spot on OKX, short perp on Binance) or you're accepting funding-rate exposure if you hold the futures leg past the funding window. A near-12% gap between a major spot venue and a major futures venue for the same asset is unusual enough that it's worth double-checking whether this is a genuine mispricing or a funding-rate-driven basis blowout — those aren't arbitraged out the same way. Executable, but this one wants a basis trader's playbook, not a pure spot-spot bot.
Buy side: Binance Futures at $0.069030. Sell side: OKX at $0.077240. Notice this is the mirror image of the BEAT trade above — buy futures, sell spot, rather than buy spot, sell futures. Seeing both directions show up on the same day between the same two venues is a decent tell that these aren't systemic mispricings so much as asset-specific noise: LAB and BEAT are each drifting independently between OKX and Binance Futures rather than the whole OKX/Binance complex being out of sync. Same infrastructure advantage as the BEAT trade (two top-tier venues), same basis-trade caveat. If you're running both LAB and BEAT simultaneously in opposite directions, you're partially netting your cross-exchange settlement risk, which is a small silver lining.
Buy side: Binance Futures at $0.072238. Sell side: Exchange24 at $0.079134. The second 龙虾/Exchange24 print of the day, this time sourcing the buy leg from Binance Futures instead of Gate. The consistency of Exchange24 sitting on the rich side of every 龙虾 pairing today reinforces the read from opportunity #1 — this looks like a venue-specific pricing lag on Exchange24 rather than random noise. Buying on Binance Futures at least solves half the liquidity-confidence problem versus buying on Gate; the open question is still what happens on the sell side once you hit Exchange24's actual resting bids. Reasonably executable in small size, same depth caveat applies.
Buy side: Bitget at $0.103772. Sell side: Exchange24 at $0.109206. Third and final 龙虾 print in the top five, and the pattern is now unmistakable: Exchange24 was rich against Gate, Binance Futures, and Bitget within the same window, at three different price levels ($0.1147, $0.0791, $0.1092 on the sell side). That's not consistent with one stale snapshot — either Exchange24 genuinely had persistently elevated demand for 龙虾 today, or there's a listing/liquidity event on that venue worth investigating directly rather than trading blind. My honest take: this is the single most interesting pattern on the board, and it's worth five minutes checking Exchange24's 龙虾 order book directly before routing anything, because if it's real and it's structural, it'll still be there in an hour; if it's a stale-quote artifact, you'll find out in about ten seconds.
The clearest structural signal in today's data is Exchange24 acting as a persistent premium venue against three different counterparties for the same asset (龙虾), at spreads ranging from 6.86% up through 16.67%. When one exchange shows up on the rich side of multiple pairings for the same ticker within a single session, it's usually one of three things: a recent listing still finding its price, thin resting liquidity that hasn't been arbed down yet by bigger players, or a data-feed/API lag on that venue's reported book. Any of the three is actionable, but each calls for a different approach — the first two reward speed, the third means the 'spread' might not be tradeable at all.
The OKX vs Binance Futures axis showed up twice today (BEAT and LAB), running in opposite directions on each asset — a sign of asset-specific dislocation rather than a systemic OKX/Binance pricing gap. The Bybit vs Binance Futures pairing also appeared twice, on LSK (7.90%) and Q (7.50%), both with Binance Futures on the buy side and Bybit on the sell side — that directional consistency is worth watching tomorrow, since a repeated pattern across unrelated assets can hint at a temporary funding or basis skew on one of the two venues rather than pure noise. No Hyperliquid or other DEX-perp prints showed up in today's top-tier data, so the action was entirely CEX-to-CEX today.
The board splits cleanly into two regimes. The wide prints — 龙虾 at 16.67% and 9-10%, BEAT and LAB near 12% — are the ones where speed matters most: wide spreads on thin books close fast, either because a market maker catches up or because the first arb bot through the door eats the visible depth in one clip. If you're chasing these, size small, fill fast, and don't expect the second or third clip at the same price. The narrower prints in the 6-8% range (LSK, Q, the smaller 龙虾 prints) tend to persist longer because they're closer to what a rational spread looks like after fees and transfer costs, which paradoxically makes them easier to size up into without moving the market against yourself.
Slippage is the silent tax on every number in this report — none of these percentages account for walking the book. On a token trading at $0.07-0.12 with unreported volume, even a few hundred dollars of size can move the effective fill price meaningfully, especially on the sell side into whichever venue is thinner. Position sizing rule of thumb for today's board: start with a probe clip (5-10% of intended size), confirm the realized spread after slippage matches expectations, then scale into the remainder only if the probe fill came in close to the quoted price. Don't full-size a first fill on any of these, especially the Exchange24 pairings where book depth is a total unknown.
Walking through the top opportunity as a worked example: 龙虾 at a 16.67% gross spread between Gate Futures and Exchange24. Strip out trading fees first — assume a blended ~0.05% taker fee on the Gate Futures leg and ~0.10-0.15% taker fee on the Exchange24 spot leg (smaller/exotic venues frequently run standard-tier retail fees rather than VIP futures rates), for roughly 0.15-0.20% in combined trading costs. That alone barely dents a 16.67% headline spread, leaving ~16.4-16.5% before transfer costs.
The bigger bite is withdrawal and settlement. If you're not already holding pre-positioned balances on both venues, moving 龙虾 from Gate to Exchange24 (or stablecoin collateral the other way) costs a flat withdrawal fee plus a confirmation window — on many chains that's 10-30 minutes of price exposure during which the spread can compress or invert entirely. On a small clip, a flat withdrawal fee can eat 1-3% of notional; on a larger clip it shrinks toward irrelevance, which is exactly why pre-funding both exchanges and skipping the live transfer is the standard professional move for repeat arb on the same pair. With balances pre-positioned, net profit on this trade realistically lands in the 15-16% range even after fees — genuinely attractive if the book depth holds up, which is the one thing this data set can't confirm.
For the mid-tier prints (BEAT, LAB, LSK, Q in the 7-12% range), the same math compresses fast: ~0.15-0.20% trading fees plus a funding-rate consideration on the futures leg, and — critically — these are spot/futures pairings, so 'withdrawal' isn't the friction, funding cost and margin requirements are. Minimum spread worth chasing on a clean spot-spot pair with pre-funded accounts: roughly 0.5-1.0% after fees, which is the standard professional threshold. On this data set, given the unconfirmed liquidity, I'd personally raise that bar to 3%+ net before committing size, simply because the $0.0M volume readout means you're pricing execution risk blind.
Watch 龙虾/Exchange24 first thing tomorrow — if that venue's book is still detached from Gate, Binance Futures, and Bitget at the open, this is a structural mispricing worth building a standing bot for rather than a one-day fluke. Keep an eye on the OKX/Binance Futures axis too; BEAT and LAB both ran hot there today in opposite directions, and a third asset showing the same pattern would confirm something systemic rather than asset-specific noise. The Bybit/Binance Futures pairing (LSK, Q) is worth a standing watch as well given the repeated directionality — buy Binance Futures, sell Bybit showed up twice today on unrelated tickers. Best windows to watch based on today's clustering: right around major-venue funding settlement times, when basis trades tend to widen briefly before arbing back closed. Prioritize confirming real order-book depth on Exchange24 and Gate Futures before tomorrow's session opens — that's the single biggest unknown hanging over today's entire board.
Sixty-five windows, one standout pattern, and a volume feed that's telling us to check before we leap. The spreads are real on the screen — whether they're real in the fill is homework for tomorrow's desk. Trade the depth you can confirm, not the number on the page.
Arbitrage Hunter — September 12, 2026