Arbitrage Trading Bot: 8 Platforms Tested & Honest Verdict (2026)

Arbitrage trading bot software promises risk-free profit: buy low on one venue, sell high on another, pocket the spread. The pitch is decades old. The reality in 2026 is that professional market makers, colocated servers, and exchange fee schedules compress most visible spreads before a retail bot can act — and the platforms selling "arbitrage bots" charge monthly subscriptions whether you capture a single basis point or not.

This guide is the honest ranking inside the broader trading bot for crypto landscape. We tested eight arbitrage trading bot platforms — Coinrule, WunderTrading, Bitsgap, 3Commas, Cryptohopper, Pionex, HaasOnline, and Pionex's triangular mode — on the things that actually decide outcomes: fee models, exchange coverage, latency assumptions, and whether the spread math survives taker fees on both legs. Then we explain the framing competitors skip: arbitrage bots solve spread capture, not asset selection. If you want profitable entries on Solana, you still need the right wallet — which is exactly where copy trading the best Solana traders changes the equation.

An arbitrage trading bot is automated software that monitors prices across two or more trading venues and executes simultaneous or near-simultaneous trades when the same asset trades at different prices. The bot's job is speed and discipline: detect the spread, size the trade, execute both legs before the gap closes, and log the result.

Arbitrage is not a directional bet. You are not predicting that BTC goes up. You are betting that BTC on Binance is momentarily cheaper than BTC on Bybit — and that you can buy and sell fast enough to lock the difference after fees.

• Cross-exchange arbitrage. Buy ETH on Exchange A at $3,400, sell ETH on Exchange B at $3,405. The $5 spread must exceed taker fees on both legs plus withdrawal/deposit costs if you move inventory between venues.
• Triangular arbitrage. Exploit pricing inconsistencies within one exchange across three pairs — e.g., BTC/USDT → ETH/BTC → ETH/USDT. No withdrawal needed, but competition is fierce and spreads are tiny.
• Funding-rate arbitrage. Long spot + short perpetual (or vice versa) to capture the funding payment when perp trades at a premium or discount to spot. This is carry, not instant spread — positions stay open for hours or days.
• DEX arbitrage. Route swaps across multiple Solana or EVM pools when the same token prices diverge. Requires sub-second execution and competes directly with MEV searchers and professional market makers.

Every type shares the same constraint: the spread must beat all costs on both legs. Retail bots that ignore withdrawal fees, taker rates, and slippage on thin pairs report paper profits that evaporate live.

The mechanics look simple in a diagram. Live execution is where retail accounts bleed.

1. Price feed aggregation. The bot connects to multiple exchange APIs (REST or WebSocket) and maintains a real-time order book for each venue. Latency here matters — a 200ms delay means the spread you saw may already be gone.
2. Spread detection. The bot calculates net spread after estimated fees: (sell_price - buy_price) - (buy_taker_fee + sell_taker_fee) - slippage_buffer. Only trades above a minimum threshold trigger.
3. Execution. Ideally both legs fire simultaneously. In practice, one leg often fills first — leaving you with inventory risk if the second leg fails or prices move.
4. Inventory management. Cross-exchange arb requires pre-funded accounts on both venues. Triangular arb keeps everything on one exchange but needs sufficient balance in all three assets.
5. Logging and alerts. Serious platforms log every attempted arb, fill rate, and net PnL. Most affiliate listicles skip this step entirely.

Professional arbitrage desks run colocated servers next to exchange matching engines, pay for direct market access, and employ engineers who optimize execution paths full-time. A $29/mo SaaS bot running on a home internet connection is not playing the same game — it is playing leftovers.

Search results mix four distinct strategies under one label. Each has different capital requirements, risk profile, and realistic profit ceiling for retail.

Arbitrage type	Capital need	Speed requirement	Retail viability
Cross-exchange spot	High (funds on 2+ CEXes)	Medium	Low — spreads rarely beat fees
Triangular (single CEX)	Medium	High	Very low — HFT dominated
Funding-rate carry	High	Low	Moderate — if you size hedges correctly
DEX / on-chain	Medium	Extreme	Very low — MEV competition
Statistical arb	High	Medium	Low — requires quant infrastructure

Funding-rate arbitrage is the one subtype where retail traders occasionally find sustainable edge — but it is a carry trade, not the instant "buy low sell high" fantasy. You hold a delta-neutral position for days, collect funding payments, and manage liquidation risk on the perp leg. Platforms like Hyperliquid and Binance futures expose these rates publicly; the hard part is sizing and rebalancing, not finding the opportunity.

DEX arbitrage on Solana routes through Jupiter, Raydium, and Orca pools. The spreads exist for milliseconds. Competing against Jito bundle searchers with dedicated RPC infrastructure is not a realistic retail strategy in 2026 — our Solana MEV guide explains why.

Ranked by fee model, arbitrage mode coverage, exchange integrations, and how honestly each platform handles the "spread vs fees" reality. Pricing confirmed on each vendor site as of August 2026 — verify before subscribing.

Platform	Pricing	Arb modes	Exchanges	Best for
Coinrule	~$30–450/mo	Cross-exchange rules, spread alerts	10+ CEXes	No-code arb rule builder
WunderTrading	~$9–49/mo	Arbitrage bot template	10+ CEXes	Budget cross-exchange monitoring
Bitsgap	$29–149/mo	COMBO arbitrage mode	16+ CEXes	Multi-exchange arb dashboard
3Commas	~$22–59/mo	Smart trade arb triggers	15+ CEXes	Custom arb + safety orders
Cryptohopper	~$19–99/mo	Arbitrage strategy templates	12+ CEXes	Arb + signal marketplace combo
Pionex	0.05%/side (no sub)	Triangular arb bot (built-in)	Pionex only	Triangular arb on one exchange
HaasOnline	~$15–75/mo	Custom HaasScript arb	20+ CEXes	Developers writing arb logic
Crypto.com bot	Exchange fees only	Perp-spot arb (eligible users)	Crypto.com only	Funding-rate carry on one venue

Coinrule — best no-code arbitrage trading bot {#coinrule-arb}

Coinrule markets a dedicated arbitrage bot page — and the product is genuinely useful for what it actually does: monitor spreads across Binance, Kraken, and Bybit, alert when a threshold is breached, and execute rule-based trades with sizing limits and timeouts. It is not HFT. It is disciplined spread checking with logs.

Trade-offs: subscription starts around $30/mo for meaningful exchange coverage, execution is not colocated, and the spreads that trigger alerts are often the ones every other retail bot already saw. Best fit for traders who want structured monitoring without writing Python. Not a fit for Solana on-chain arb.

WunderTrading — budget arbitrage trading bot {#wundertrading-arb}

WunderTrading ships an arbitrage bot template at lower subscription tiers than Bitsgap or 3Commas. Connect exchange APIs, set minimum spread thresholds, and let the bot attempt cross-exchange fills. The UI is clean and the pricing is accessible for experimentation.

The same structural limits apply: CEX API keys, no Solana DEX routing, and spreads that look profitable on screen often vanish after taker fees on both legs. Useful for learning how arb math works; unlikely to generate meaningful returns on a $3k account after subscription costs.

Bitsgap — best multi-exchange arbitrage dashboard {#bitsgap-arb}

Bitsgap COMBO mode is the platform SERP results reference most often for crypto arbitrage. It monitors prices across 16 exchanges, highlights spread opportunities, and can execute paired trades from one dashboard. Demo mode lets you paper-trade arb without risking capital — genuinely valuable for understanding fee impact.

Subscription math is the catch. At $69/mo Advanced on a $5k account, you pay 16.6% APR in fixed costs before the bot captures a single spread. Most visible cross-exchange spreads on BTC and ETH are under 0.1% — and taker fees on two CEX legs eat 0.15–0.20%. The math only works on alt pairs with wider spreads, which carry higher slippage and inventory risk.

3Commas, Cryptohopper, Pionex, HaasOnline {#other-arb-platforms}

3Commas does not ship a dedicated "arbitrage bot" button, but smart trades with paired exchange triggers achieve the same outcome — buy on Exchange A when spread exceeds X, sell on Exchange B. More configurable than WunderTrading, same CEX-only blind spot.

Cryptohopper lists arbitrage among its strategy templates and pairs it with the signal marketplace. Useful if you want arb monitoring plus directional signals; adds complexity without solving the fee problem.

Pionex is the outlier with a built-in triangular arbitrage bot at 0.05% per side — no subscription. The bot cycles through three pairs on Pionex's own order book. Competition is intense and spreads are microscopic, but the fee model beats subscription platforms for small accounts experimenting with triangular logic.

HaasOnline targets developers who write custom arbitrage scripts in HaasScript. Full control, 20+ exchange integrations, but you need engineering skill and a VPS. Same category as self-hosted Gunbot — powerful if you can build the strategy, expensive in time if you cannot.

Three cost layers decide whether an arbitrage trading bot is net-positive on your account size:

• Platform cost. Subscription ($108–$1,788/yr on most platforms) or per-trade (Pionex 0.05%/side). On a $5k account, a $30/mo subscription is 7.2% APR drag before a single arb fill.
• Exchange taker fees. Cross-exchange arb pays taker on both legs. Binance spot taker at 0.10% × 2 legs = 0.20% minimum spread required just to break even — before slippage, withdrawal fees, or transfer time.
• Inventory and transfer costs. Moving USDT between exchanges costs withdrawal fees ($1–$25 depending on chain) and takes 10–60 minutes. The spread you saw when you initiated the transfer may not exist when funds arrive.

Example: $10k split across Binance and Bybit, Bitsgap Pro ($42/mo), monitoring BTC/USDT cross-exchange spreads. Average visible spread when alerts fire: 0.08%. Taker fees both legs: 0.20%. Net per attempted arb: -0.12% before slippage. Subscription drag: 5% APR. Result: negative expected value on every fill, compounded by the subscription.

Triangular arb on Pionex avoids withdrawal costs but competes against the exchange's own market makers and every other Pionex user running the same bot. Spreads that survive 0.10% round-trip fees are rare on major pairs.

This is the same structural issue we document in our is copy trading profitable analysis and our grid bot trading fee math: fixed-cost bots punish small accounts regardless of strategy quality.

Five failure modes that consume retail arbitrage accounts. None appear in affiliate listicles because admitting them undermines the "risk-free profit" pitch.

• Spread compression. BTC/ETH spreads across top CEXes rarely exceed 0.05% for more than a few seconds. Professional market makers with colocated infrastructure capture these before your API call returns.
• Leg risk. You buy on Exchange A, but the sell on Exchange B fails or fills at a worse price. You are now holding directional exposure you did not plan for — the opposite of "risk-free" arbitrage.
• Withdrawal delays. Cross-exchange arb without pre-funded accounts requires transfers. Transfer time = spread risk. Most profitable arbs require capital already sitting on both venues.
• Alt-pair illusion. Wider spreads on thin alt pairs look attractive until slippage eats 0.5–2% per leg. The bot logs a "successful" arb that lost money net of slippage.
• Regulatory and access risk. Exchange API access can be revoked, withdrawal limits can block rebalancing, and geographic restrictions can strand inventory on one venue.

Funding-rate arbitrage avoids some of these — you stay on one exchange — but introduces liquidation risk on the perp leg, basis risk if the funding rate flips, and capital lockup that prevents deploying the same funds elsewhere.

Solana DEX arbitrage is a different game from CEX cross-exchange arb — and a harder one for retail. Prices across Raydium, Orca, Meteora, and Jupiter routes diverge constantly as pools rebalance. The opportunity window is measured in slots, not seconds.

Professional searchers run dedicated RPC nodes (Helius, Triton, QuickNode gRPC streams), submit Jito bundles with priority tips, and compete in sub-400ms auctions for transaction ordering. Building or buying an off-the-shelf "Solana arbitrage bot" without that infrastructure is like bringing a bicycle to a Formula 1 race.

Our Solana MEV guide documents the extraction mechanics in detail. For most Solana traders, the realistic path is not building an arb bot — it is using a copy trading bot that mirrors wallets already capturing on-chain opportunities, with sub-400ms execution via uwuu.ai and performance-based fees that only charge when you profit.

These are fundamentally different bets. Arbitrage bots compete for microscopic spreads against professional infrastructure. Copy trading competes against nobody — you mirror a wallet that has already demonstrated profitable timing.

Dimension	Arbitrage trading bot	Copy trading bot
Edge source	Price discrepancies across venues	Another wallet's proven skill
Competition	HFT firms, MEV searchers, market makers	Front-running risk only
Infrastructure	Colocated servers, dedicated RPC, Jito bundles	Wallet connection + copy key
Fee model	Subscription + exchange taker fees	Performance-based (pay on profit only)
Solana support	Requires custom on-chain infra	Native (uwuu, GMGN, etc.)
Retail viability 2026	Very low on spot/DEX arb	Moderate with right trader selection

If you have $50k+ split across multiple CEXes, engineering skill, and a VPS next to exchange servers, funding-rate carry and occasional cross-exchange arbs can add incremental yield. If you have a $3k Solana wallet and want automated exposure to profitable trades, copy trading from the uwuu leaderboard is the path that matches your infrastructure. See how to copy trade on Solana for the setup walkthrough.

If you still want to experiment with CEX arbitrage after reading the fee math above, follow this checklist to limit damage:

1. Start with paper trading. Bitsgap demo mode and Coinrule's sandbox let you log arb attempts without risking capital. Run for two weeks and calculate net PnL after simulated fees.
2. Pre-fund both venues. Never arb with live transfers. Keep equal USDT balances on both exchanges so both legs can fire instantly.
3. Set minimum spread thresholds above fee breakeven. If your combined taker rate is 0.20%, set minimum spread to 0.30%+ to leave room for slippage.
4. Cap position size. Never risk more than 2–5% of total capital per arb attempt. Leg risk on a failed second fill can turn a "risk-free" trade into a directional loss.
5. Log every fill. Export trade logs weekly. If net PnL after all fees is negative over 30 days, stop the bot and reallocate capital — subscription costs do not pause while you debug.
6. Compare against copy trading. Run the same capital through a Solana copy trading bot for one month and compare net returns. Most retail traders find copy trading outperforms arb after accounting for subscription drag.

Does crypto arbitrage still work in 2026?

Professional desks still profit from funding-rate carry, statistical arb, and DEX routing — but retail cross-exchange spot arbitrage on major pairs is largely compressed by market makers and fees. Triangular arb on a single exchange survives only on thin alt pairs with higher slippage risk. For most retail traders, the edge has moved to strategies that do not compete on milliseconds — like copy trading proven wallets.

What is the best arbitrage trading bot?

For no-code CEX monitoring, Coinrule offers the cleanest spread-alert workflow. For multi-exchange dashboards, Bitsgap COMBO mode is the most polished. For triangular arb without a subscription, Pionex's built-in bot at 0.05% per side is the best fee model for small accounts. None of these solve the fundamental problem that visible spreads rarely beat taker fees on both legs.

Are arbitrage bots illegal?

Arbitrage trading itself is legal in most jurisdictions — you are buying and selling the same asset at different prices, not manipulating markets. However, using bots may violate specific exchange terms of service, and some jurisdictions restrict automated trading on certain platforms. Always check your exchange's API terms and local regulations before deploying an arbitrage trading bot.

Is AI arbitrage profitable?

"AI arbitrage" marketing usually means rule-based spread detection with a machine-learning label — not genuine predictive AI. The profitability question is the same as any arb bot: does the spread exceed all fees on both legs? In 2026, AI branding does not change the math. Funding-rate carry with proper hedging is the one arb subtype where retail occasionally finds sustainable returns, but it is a carry trade, not instant spread capture.

Can I run an arbitrage trading bot on Solana?

Technically yes — Jupiter and other aggregators expose routing APIs, and open-source Solana arb bots exist on GitHub. Practically, you compete against MEV searchers with dedicated RPC infrastructure and Jito bundle access. For most Solana traders, copy trading via uwuu.ai delivers better risk-adjusted returns without building arb infrastructure from scratch.

How much capital do I need for an arbitrage trading bot?

Cross-exchange arb requires meaningful balances on at least two venues — realistically $10k+ split across exchanges to make 0.05–0.15% spreads worth the operational overhead. Triangular arb on Pionex can start smaller ($1k+) but faces the same spread compression. Factor in subscription costs: a $30/mo bot on a $5k account costs 7.2% APR before any trades execute.

Arbitrage trading bot software will always sound safer than directional trading — because "risk-free spread capture" is a compelling story. In 2026, the spreads that survive professional competition and retail fee schedules are thinner than the marketing suggests. Funding-rate carry works for capitalized, patient traders. Everything else is increasingly an infrastructure game that subscription SaaS bots cannot win.

If your goal is automated Solana exposure without building arb infrastructure, start with the uwuu leaderboard: pick a trader with verified on-chain PnL, connect your wallet, and let copy trading handle execution. Performance-based fees mean you only pay when the strategy works — which is more than most arbitrage bots can honestly claim.