177 cross-exchange arbitrage opportunities crossed the desk today, headlined by a 49.34% OPG spread between Binance Futures and Gate Futures and a cluster of near-49% T spreads on the same pair. Gate-to-Binance Futures dislocations dominated the board, with OKX showing up as the consistent sell-side venue against Bitget and Binance on the COHR and WAL pairs.
177 flagged opportunities crossed the tape today, and the top of the board was ugly in the best possible way. OPG printed a 49.34% spread between Binance Futures ($0.105600) and Gate Futures ($0.157700) — the kind of number that makes you check your data feed twice before you check your margin balance. Spreads north of 45% don't happen on deep, efficiently-arbed majors. They happen on thin-book alts where two venues are quoting off different reference points, funding is diverging, or one exchange's futures market simply isn't getting arbed by size yet. That's exactly the profile we saw across the board today.
The session had a clear signature: Gate Futures pricing T meaningfully below Binance Futures, repeatedly, across five separate snapshots ranging from 46.79% to 48.90%. That's not noise — that's a structural pricing gap that persisted long enough to get sampled five times. Layer on WAL (39.13%, Binance Futures to OKX) and a trio of COHR prints in the 35.6%–35.9% range (Bitget/Binance Futures to OKX), and you've got a session where three distinct assets each threw off double-digit, repeatable dislocations. For an arb desk, that's either a very good day or a very suspicious one — often both at once.
One important caveat before we get into the trade-by-trade breakdown: our feed did not capture executable order-book depth or fill volume for this session (pump/dump/buy/sell pressure all read $0.0M). Everything below on size and window duration is analyst-estimated from typical liquidity profiles for these pairs, not a confirmed fill. Treat it as a starting assumption to verify live, not a guarantee of size.
Buy OPG Futures on Binance at $0.105600, sell (short) on Gate Futures at $0.157700. This is the widest print of the session by a comfortable margin. At this magnitude, the trade isn't really about fee optimization — it's about whether you can get filled at anything close to the quoted prices before the book moves. Spreads this size on a low-priced alt (sub-$0.11) almost always mean thin depth on at least one leg; Gate Futures books on smaller-cap perps routinely can't absorb more than a few thousand dollars of size without 5-10% of slippage eating into the spread. Our take: executable in small clips (think $500-$2,000 per leg, scaled in), not as a single block trade. If you tried to push $50k through this in one shot, you'd likely close most of the spread yourself just from market impact. Watch funding rates on both legs closely — a 49% price gap on two live futures markets usually comes with a wild funding differential that can either pad your return or eat it alive depending on which side you're carrying.
Buy T Futures on Gate at $0.002431, sell on Binance Futures at $0.003620. This is the first of five T prints clustered in the same range, which tells us this wasn't a one-tick fluke — Gate was persistently underpricing T relative to Binance for a meaningful stretch. Sub-$0.003 tokens on Gate Futures tend to have shallower books than their Binance counterparts, so the realistic read is that Binance's futures market was the more liquid, better-arbed venue and Gate simply lagged. Executable, but size-limited — this is a scale-in trade with tight per-clip caps, not a single large fill. The repetition across five snapshots is actually the more valuable signal here: it suggests a recurring, possibly systematic gap rather than a one-off wick, which is worth setting a standing monitor on rather than chasing once.
Buy T Futures on Gate at $0.002463, sell on Binance Futures at $0.003630. Second print in the T cluster, roughly 30 minutes' worth of price drift from the first (Gate's price ticked up from $0.002431 to $0.002463 while Binance held closer to flat). The persistence across snapshots is the story: this wasn't a single spike that got arbed away in seconds, it was a gap that kept re-forming. That's a much better setup for an arb desk than a single flash print, because it gives you multiple entry windows instead of one blink-and-you-missed-it moment. Risk is still liquidity-bound on the Gate leg — treat any T Futures fill above low four-figures on Gate with caution until you've confirmed the book depth live.
Buy T Futures on Gate at $0.002446, sell on Binance Futures at $0.003622. Third consecutive T print, spread holding just above 48%. At this point the pattern is unmistakable — Gate's T Futures market was running a persistent 45-49% discount to Binance across the whole session window. For arb purposes, this is the most interesting entry of the five: the spread was neither at its widest (that was print #2) nor showing signs of imminent collapse, which historically means there was still runway left to work size into the position before convergence. Still capped by Gate-side depth, but a trader who caught this mid-cluster rather than chasing the first print likely got a cleaner average fill.
Buy T Futures on Gate at $0.002453, sell on Binance Futures at $0.003622. Fourth of five T snapshots, spread compressing slightly from print #4 but still enormous by any normal standard. The Binance-side price barely moved across all four T prints (~$0.003620-$0.003630), which is the tell that Binance Futures was the anchor — the efficient, well-arbed market — while Gate drifted underneath it. Late entries into a persistent-gap trade like this carry more convergence risk than early ones: you're betting the mispricing holds a little longer, not that it's fresh. Still worth a scaled clip, but size down relative to prints #2-4.
The dominant pattern today was Gate Futures pricing below Binance Futures on lower-cap perps — that's the entire T cluster (five prints, 46.79%-48.90%) and it's consistent with what we typically see on Gate: solid listing coverage on smaller alts, but thinner futures depth than Binance, which lets price gaps open and persist longer before size arbs them shut. The second pattern was OKX sitting on the expensive side of the trade against both Bitget and Binance Futures — visible on WAL (Binance Futures cheap, OKX expensive, 39.13%) and all three COHR prints (Bitget and Binance Futures cheap, OKX expensive, 35.65%-35.90%). If you're building a standing watchlist, 'OKX runs rich relative to Binance/Bitget on lower-liquidity names' is the pattern worth automating an alert around — it showed up on two unrelated assets in the same session, which is more than coincidence.
The classic arb tradeoff is on full display today: the widest spreads (OPG at 49.34%, the T cluster at 46-49%) are on the lowest-liquidity legs, meaning you can move fast and small, but you cannot move fast and large. Slippage on a $0.0025-ish token's futures book scales brutally once you're past a few thousand dollars per clip — a market order that looks fine on a $500 test size can easily eat 2-4% of your edge once you triple it. The rule of thumb we'd apply here: for any spread above 30%, assume the quoted price is only good for the first 10-20% of whatever size you're planning, and scale in with limit orders rather than sweeping the book with market orders on either leg.
Position sizing recommendation: split your intended notional into 4-6 clips, execute both legs of each clip near-simultaneously (use a script or at minimum two browser tabs with orders staged), and reassess the spread after each clip rather than committing the full size upfront. On the T cluster specifically, the fact that we saw five separate snapshots of a similar-magnitude gap is actually good news for size — it suggests the mispricing had some duration to it, so a trader who missed the first print still had three or four more entry windows rather than one. That's a materially lower-risk way to build size than chasing a single flash print on OPG or WAL.
Walking the OPG trade at a modest $3,000-per-leg size: buy $3,000 of OPG Futures on Binance at $0.105600 (≈28,409 OPG), simultaneously short $3,000 notional on Gate Futures at $0.157700. Gross spread capture on full convergence: 49.34% of $3,000 ≈ $1,480. Trading fees: assume 0.05% taker per fill on futures, four fills total (open Binance long, open Gate short, close Binance, close Gate) = 0.20% of notional ≈ $6. If you're pre-positioned with margin/collateral on both exchanges (which you should be for any trade this fast-moving — there's no time to wait on a withdrawal), there's no on-chain withdrawal fee to subtract on the trade itself. Net: roughly $1,474, or a 49.1% net spread. At this magnitude, fees are rounding error — the entire risk is execution and slippage, not cost.
Now the more realistic, everyday case — a typical 0.6% cross-exchange spread on a liquid pair, $10,000 per leg: gross profit ≈ $60. Fees at 0.05% taker × 4 fills = 0.20% ≈ $20. If the trade does require moving USDT between exchanges (rather than running pre-funded dual accounts), add a network transfer: TRC20 USDT withdrawal runs roughly $1, ERC20 can run $2-8 depending on gas — call it $3 average. Net profit: ≈$37 on $10,000 deployed, or about 0.37% net. That's the real-world math most arb desks live in day to day, and it's why pre-funding both exchanges (avoiding the withdrawal leg entirely) roughly doubles your realistic take-home on bread-and-butter spreads.
Minimum spread worth chasing: with a 0.20% round-trip fee floor (four taker fills) plus execution slippage on anything beyond top-of-book size, we'd set the practical floor at 0.5%-0.6% gross spread for it to be worth the operational risk on a manually-executed trade, and closer to 0.25%-0.3% if you're running maker orders and have both exchanges pre-funded with automated execution. Everything in today's top 10 — 35.65% to 49.34% — clears that bar by an order of magnitude, which is exactly why liquidity and execution risk, not fees, are the binding constraint on these particular trades.
Keep the Gate Futures vs Binance Futures pair on a tight watch for T — five recurring prints in one session on the same pair is a strong signal this gap isn't fully closed, and it's the kind of thing that reappears the next time volatility ticks up on the name. Same logic applies to OKX vs Binance/Bitget on WAL and COHR: two unrelated assets showing the same directional bias (OKX rich, Binance/Bitget cheap) in one session is worth an automated spread alert rather than a one-off check. Best windows to watch: the first hour after Asia-session open and the first hour of US trading, when funding resets and liquidity providers reposition tend to create the transient mispricings that show up as our top-of-board spreads. OPG is a wildcard — a single 49% print doesn't establish a pattern the way the T cluster does, so treat any repeat appearance tomorrow as the confirmation signal, not the entry itself.
177 spreads on the board, five of them on the same T pair alone — that's not a quiet session, that's a market telling you where the liquidity gaps are. Size small on the wide ones, watch the recurring ones, and don't let a 49% number make you skip your due diligence on depth. Fees won't kill these trades. Sloppy execution will.
Arbitrage Hunter — July 22, 2026