Cheapest Bridge to Arbitrum

Picking the right bridge to Arbitrum

• Official Arbitrum Bridge. Use bridge.arbitrum.io. Zero protocol fee, ~92k mainnet gas. Cheapest in absolute terms for transfers above ~$500.
• Across or Stargate (paid). Delivers in seconds via L2 liquidity. 0.05–0.3% protocol fee plus L2 gas. Beat the official bridge on small transfers ($50–$300 range) where mainnet gas dominates.
• Exchange direct withdrawal. Binance, OKX, and some others support direct USDC / USDT / ETH withdrawal to Arbitrum. Skips bridging entirely — cheapest if your funds are already on a CEX.

Why Arbitrum's native bridge is cheaper than Base or Optimism

All three major L2 native bridges work the same way conceptually: lock funds in an L1 contract, emit an event, have the L2 sequencer mint the equivalent on the destination side. The cost difference comes down to L1 contract architecture:

• Arbitrum: ~92,000 gas. The Arbitrum Inbox contract is the leanest of the three. The deposit transaction is essentially “lock ETH, write a retryable ticket, emit event.” The retryable ticket mechanism is what makes Arbitrum's bridge unique — it queues the L2 execution for the sequencer to pick up, with built-in retry logic if something fails on the L2 side.
• Base / Optimism (OP Stack): ~130,000 gas. The OptimismPortal contract does similar work but with more complex event encoding (the cross-domain message format includes more metadata) and tighter coupling with the L1 state tracking required for fault proofs.
• Polygon PoS: ~80,000 gas. The lowest of all, but Polygon isn't comparable — it's a sidechain, not a rollup. The deposit contract is simpler because it doesn't need to track L1 data for security.

At 20 gwei mainnet gas and ETH ≈ $3,500, the Arbitrum deposit costs about $6.50; the same deposit to Base costs about $9. That ~30% difference compounds on every deposit and is the main reason Arbitrum often wins the “cheapest bridge” question for large transfers, even when Base has cheaper L2 gas per transaction afterwards.

How a retryable ticket works (and why it matters)

Arbitrum's retryable ticket is the L1→L2 message that the Inbox contract creates. It bundles three things: your deposit amount, the L2 address to credit, and an optional piece of calldata to execute as part of the credit. The sequencer picks up the ticket and executes it on L2, typically within minutes.

Two consequences worth knowing:

• Tickets can expire. Retryable tickets have a built-in lifetime (~7 days for the auto-redeem; up to 30 days total before they're garbage-collected). In normal operation this doesn't matter — the sequencer redeems them immediately. If the sequencer is offline for an extended period, or if there's a problem with the L2-side execution, you have a window to retry manually.
• Retry costs L2 gas. If the auto-redeem fails (rare, but happens with custom token contracts or contracts that revert during the deposit step), you can call redeem() manually on the Arbitrum side. This costs L2 gas — usually a few cents — but it's not free. Watch for this if you're bridging exotic tokens.

What happens when the Arbitrum sequencer goes down

Arbitrum currently runs a single centralized sequencer operated by Offchain Labs. It has gone down a handful of times since launch, usually for 30-90 minutes. During an outage, three things happen — relevant if you're mid-bridge:

• Your L1 deposit confirms normally. The mainnet side doesn't care about Arbitrum's sequencer state. Your funds are locked in the Inbox contract.
• L2 credit is delayed. The sequencer can't redeem your retryable ticket while it's offline. Your funds aren't lost — they're just stuck in transit until the sequencer recovers. Usually 30-90 min; rarely longer.
• There's an emergency exit. If the sequencer is offline for >24 hours, you can force-include the L2 transaction through L1 directly. It's slow and costs mainnet gas, but it works. Not relevant for routine outages.

If you're bridging large amounts and want sequencer-independence, the paid bridges (Across, Stargate) don't depend on the Arbitrum sequencer the same way — they settle on their own infrastructure. Different tradeoff: you trust the paid bridge's multisig instead of the sequencer's uptime.

Frequently asked questions