Thirty-three arbitrage opportunities crossed the board today, led by a 20.14% MMT spread between Binance and Bybit Spot. Uncle Sol breaks down the top five prints, the exchange-pair patterns behind them, and the real fee-adjusted profit after trading costs, withdrawal drag, and slippage.
Uncle Sol back at the terminal, and today's scan came back loud. Thirty-three arbitrage opportunities crossed the wire in the last session — that's a busy board by any measure, and it tells you two things at once: liquidity is fragmenting hard across venues right now, and somewhere out there a market maker's price feed is lagging. Both are good news for anyone running a cross-exchange book.
The headline number is MMT at a 20.14% spread — buy side on Binance at $0.412900, offloading on Bybit Spot at $0.427500. That's not a rounding artifact from a thin order book blip. A spread that size on a token trading in the low forty-cent range, on two of the largest venues in the business, is the kind of print that either means the market briefly forgot how to price MMT, or one of these two order books is so shallow that the sell-side print you're looking at is a single lonely ask sitting well above anything tradable in size. Either way, it's the story of the day — MMT shows up four more times in today's top ten, which tells you this token was getting repriced all session across nearly every venue that lists it.
Below the headline, GRVT put in two clean prints (13.20% and 7.88%) on the identical OKX-to-KuCoin route both times, CAP and KIOXIA each cracked double digits, and ERA rounded out the top ten with a tighter 7.75% gap. Let's break down what's tradable, what's a trap, and what it actually costs you in fees to find out.
This is the widest spread on the board: buy at $0.412900 on Binance, sell at $0.427500 on Bybit Spot. A 20%+ round trip before a single fee is subtracted puts this squarely in verify-before-you-touch-it territory rather than send-it territory. Spreads this size between two top-tier CEX order books almost never survive contact with real size — either the Bybit ask pricing this spread is only a few dollars deep and evaporates the moment a market order hits it, or MMT is genuinely being repriced faster on Binance than Bybit's book can follow, in which case the window is measured in single-digit minutes, not hours. My take: executable only in small clips, and only if you already have capital pre-positioned on both venues — if you need to withdraw and transfer to capture this, the spread will be gone before your transaction confirms. Treat this as a signal that MMT liquidity is fractured today, not as a standing invitation to size up.
Second MMT print of the session, and this one is cross-market as well as cross-exchange — buying the futures leg on Gate at $0.349400 and selling spot on Bybit at $0.364074. Mixing a futures buy with a spot sell means you're not running a clean cash-and-carry; you're carrying basis risk on top of exchange risk, and you have to close or roll the futures position separately from the spot leg. That's an extra moving part most spot-only arb setups don't model, which is probably part of why the gap held. Executable, but only for traders already running a perp book on Gate who can flatten the futures side without slippage of their own. For a spot-only trader, skip this one — the futures leg is the tail risk you don't want.
GRVT is the most normal-looking opportunity in the top five: buy on OKX at $0.246730, sell on KuCoin at $0.279290 — and it's the same route that shows up again lower on the list at 7.88%. Two prints on the identical pair in one session isn't noise; KuCoin is consistently marking GRVT rich relative to OKX today, likely because GRVT sees materially thinner order flow on KuCoin and its book is slower to react to OKX-side selling. This is the most executable trade in the top five for anyone already holding balances on both venues: clean spot-to-spot, no derivatives basis risk, no roll. Watch KuCoin's ask-side depth before committing size — a book that's slow to reprice is often also thin, and you can walk your own fill price down the spread you're trying to capture.
Buy CAP on KuCoin at $0.032350, sell on Bitget at $0.033790. Sub-four-cent tokens are where spread percentages get exaggerated by small absolute price moves — a gap of roughly $0.00144 per token reads as a big percentage precisely because the denominator is so small. That cuts both ways: your percentage return looks great, but so does your exposure to a single tick of slippage wiping out the edge. At this price level, withdrawal fees denominated in flat token units rather than a percentage can eat a disproportionate chunk of the spread if you're moving small size, so this trade wants either meaningful position size or pre-funded balances on both KuCoin and Bitget. Executable for the pre-funded crowd; marginal after fees for anyone financing a live withdrawal.
Buy on Bitget at $274.200000, sell on KuCoin at $289.020000. KIOXIA prices in the hundreds of dollars rather than fractions of a cent, which is unusual company for this list — the ticker overlaps with the real-world Japanese NAND-flash manufacturer, so this is very likely a pre-market or synthetic equity-style listing rather than a standard altcoin, and those instruments trade on thinner, less mature books than majors like MMT or GRVT. That combination — high per-unit price plus a newer or more exotic listing type — means realistic bid-ask spreads on both legs are probably wider than the headline 11.36% suggests. My take: worth a look with a small clip to confirm both books actually fill at the quoted levels before scaling up; don't assume KIOXIA behaves like a liquid, well-arbitraged large-cap just because the spread number looks clean.
Two route patterns stand out across today's full top-ten list, and both are worth watching going forward rather than treating today's board as a one-off. First: OKX-to-KuCoin showed up twice, both on GRVT, at 13.20% and 7.88%. When the same pair on the same route prints twice in one session, that's not coincidence — it's a structural lag between how fast OKX repriced GRVT and how fast KuCoin's book followed. KuCoin consistently looks like the slow venue in these mid-cap alt pairs; its books are thinner and its market makers appear less aggressive about closing gaps against OKX specifically, which makes OKX-to-KuCoin worth a standing watch on any token both venues list.
Second: Bitget shows up on both sides of the ledger — as the cheap buy leg against Bybit and OKX (MMT at 8.68% and ERA at 7.75%), and as the expensive sell leg against KuCoin (CAP at 11.88% and KIOXIA at 11.36%). That's not one consistent directional bias, which tells you Bitget's pricing lag isn't systemic across the board — it's token-by-token, tied to how thin each specific pair is on Bitget's book rather than any platform-wide latency issue. Treat Bitget as a venue where you check both directions per-token rather than assuming a fixed always-cheap or always-rich bias.
Third, and worth flagging separately: Gate Futures appeared as the buy leg twice (MMT at 18.85% and 7.48%), both times paired against a spot sell on a different CEX. Futures-to-spot arb across exchanges is a different risk profile than spot-to-spot — you're carrying basis and funding exposure, not just execution risk — so don't lump these into the same bucket as the clean spot pairs even when the headline percentage looks similar.
No Binance-vs-OKX or Hyperliquid-vs-CEX pattern showed up in today's top ten specifically, which is itself informative: today's dislocations are concentrated in the OKX / KuCoin / Bitget / Bybit / Gate cluster of venues, not in the largest-cap Binance order books or in the DEX-perp space. If you're only watching Binance and a single DEX for arb signals, you'd have missed essentially all thirty-three opportunities today.
Every spread on this board falls into one of two buckets, and mixing them up is how arb accounts bleed slowly instead of winning cleanly. Bucket one is the quick, small, high-percentage prints — MMT's 20.14% and 18.85% belong here. These are almost certainly thin-book artifacts: the percentage is big because the tradable size behind it is small. Chase these with small, pre-positioned clips only, expect the window to close in minutes, and never route a market order at size into a spread this wide without checking book depth first — you will move the price against yourself faster than the spread can pay you back.
Bucket two is the more moderate, more durable spread — GRVT's 13.20%/7.88% pair and ERA's 7.75% fit here better. These read like genuine cross-venue mispricings on tokens with enough real order flow that the gap isn't purely a stale-quote illusion. They're slower to close, which means more time to size a position, but they're also smaller in percentage terms, so total capital committed needs to be larger to generate the same dollar profit as a lucky hit on a bucket-one print.
Slippage is the variable that decides which bucket you're actually in, and you only find out by testing with size. Rule of thumb for today's board: start with a probe order at roughly 10-15% of your intended position on both legs, confirm your actual fill price against the quoted price, and only scale the remaining 85-90% if the realized spread after that probe fill is still comfortably above your fee-plus-buffer threshold. If the probe already erodes more than a third of the headline spread, that pair's real depth doesn't support your intended size — stop there, don't average in hoping the book refills.
Let's run the numbers on the MMT headline trade — buy Binance at $0.412900, sell Bybit Spot at $0.427500, gross spread 20.14%. Standard taker fees on both venues run about 0.10% per side for a typical retail tier, so a full round trip (buy taker plus sell taker) costs roughly 0.20% in trading fees alone. Use limit orders instead of market orders where the book allows it and you can often get maker rates closer to 0.02-0.08% per side, cutting round-trip cost toward 0.10-0.15% — worth the extra few seconds of setup on any spread this size.
Withdrawal cost is the wildcard, because it's charged as a flat token amount, not a percentage, and it only applies if you're physically moving MMT from Binance to Bybit rather than trading against balances you already hold on both venues. On a token priced around $0.41, even a modest flat withdrawal fee can represent a meaningful percentage of a small position — which is exactly why professional arb desks avoid withdrawal-based execution for the wide-but-fast bucket-one spreads entirely. If you are structured that way, budget roughly 0.5-2% of position value for withdrawal drag depending on the chain and current network congestion, and count the transfer confirmation time — anywhere from under a minute to twenty-plus minutes depending on chain — as pure risk exposure, since the spread can and often does close before the transfer lands.
Subtract slippage last, because it's the piece that actually varies trade to trade — for a thin micro-cap book, budget 1-3% of the spread eaten by slippage on anything beyond a small probe size. On the MMT trade specifically, that still leaves comfortable double-digit net profit if you're pre-funded and sized appropriately; it's a much thinner margin if you're financing a withdrawal in real time.
For the tighter prints like ERA's 7.75% or GRVT's 7.88%, the math is less forgiving: strip out 0.15-0.20% trading fees and 1-2% realistic slippage on a mid-cap alt book, and you're netting somewhere in the 5.5-6.5% range — still very tradeable, but it underlines why minimum-spread thresholds matter. My working floor for spot-to-spot CEX arb with pre-funded balances is about 0.5-0.7% gross — below that, combined round-trip fees alone eat most of the edge. If you have to route a withdrawal to capture the trade, move that floor up to 2-3% minimum, purely to survive fee drag plus transfer-time risk.
A few things every arb trader should be actively watching on today's board, not just filing away as background risk.
MMT is the name to watch first thing — five appearances in today's top ten, spanning Binance, Gate Futures, OKX, Bitget and Bybit, is not the footprint of a token whose market makers have converged on a single fair price yet. Expect continued dislocation, most likely reappearing on whichever pair had the largest session-end order flow imbalance; check Binance-to-Bybit and Bitget-to-OKX first.
GRVT's OKX-to-KuCoin route deserves a standing watch rather than a one-day look — it printed twice today at different points in the session, which is the signature of a recurring lag rather than a random blip. If that pattern holds, expect it to show up again on the same route within the next 24-48 hours.
Best times to watch, based on today's cluster of prints: the early Asia session open and the US-to-Asia handoff tend to be when OKX, KuCoin and Bitget order books briefly desynchronize from Binance and Bybit, since regional flow shifts faster than smaller venues' market makers can rebalance quotes. Keep an eye on Gate Futures funding resets too, since both futures-linked MMT prints today lined up with basis moves rather than pure spot dislocation.
Thirty-three cracks in the market's pricing armor today, and MMT wearing the crown at 20.14% — that's a good hunting day if you move fast, size sensibly, and don't fall in love with a headline percentage before checking what's actually behind it. Watch your fee stack, respect the withdrawal clock, and don't chase a spread you can't verify. Stay sharp out there.
Arbitrage Hunter — July 31, 2026