Base, the Ethereum layer-2 network built by Coinbase, will activate its Beryl upgrade on mainnet on June 25, 2026, at 18:00 UTC, the team confirmed in an engineering post published late on June 21. The release adds a native token standard called B20, shortens the standard withdrawal wait from Ethereum from seven days to five, and swaps in a faster node client. Beryl lands just four weeks after Base’s prior Azul upgrade, a cadence the team says was not possible before it rebuilt its software stack.
A layer 2 is a separate network built on top of Ethereum that processes transactions faster and more cheaply before settling them on Ethereum’s main chain. A token standard is a shared template that defines how a token is created and how wallets, exchanges, and apps interact with it.
Key takeaways
- Base activates the Beryl upgrade on mainnet June 25, 2026, at 18:00 UTC, with the testnet version live on Base Sepolia since June 18.
- Beryl introduces B20, a native token standard that runs as code embedded in the node software in Rust rather than as a smart contract on top of the chain.
- The upgrade cuts the single-proof withdrawal window from seven days to five and ships Reth V2, which Base says reduces node disk use by about half and raises throughput by roughly a third.
- B20 targets stablecoin and tokenized-asset issuers with built-in compliance controls, including the ability to freeze and burn tokens held at blocked addresses.
Published: June 22, 2026, 09:00 UTC
What Base is shipping with Beryl
The centerpiece is B20, which Base describes as a superset of the ERC-20 standard that powers most tokens on Ethereum. Because it implements the ERC-20 specification, a B20 token stays compatible with existing wallets, block explorers, and onchain protocols with no special integration. What changes is where the logic runs. A B20 is a precompiled contract, meaning its rules execute natively inside the node software, written in Rust, instead of running as bytecode on the Ethereum Virtual Machine. Base says that approach lowers transfer costs, reduces the data each node must store, and increases throughput compared with a conventional smart-contract token.
Base built the standard for issuers of stablecoins and real-world assets, the same category of products driving competition between layer-2 networks. B20 ships with an issuer toolkit that includes role-based permissions, optional supply caps, signature-based approvals, and transfer policies that authorize movements by sender and receiver. It also includes a function called burnBlocked, which lets an issuer burn the balance of a policy-blocked address. That feature supports the freeze-and-seize path that regulated issuers are often required to operate, a capability that has drawn scrutiny as stablecoin oversight tightens. Two B20 variants are available at launch: a general-purpose asset token with configurable decimals and rebasing, and a stablecoin token fixed at six decimals with a self-declared currency code.
Why faster withdrawals matter
Beryl shortens the time a user waits to finalize a withdrawal from Base back to Ethereum on the most common path, moving it from seven days to five. Withdrawing assets across the bridge happens in two steps, initiating and finalizing, separated by an enforced delay that lets the network catch a faulty proof before funds settle. The seven-day window dates to Base’s older fault-proof design. After the Azul upgrade introduced a multiproof system in May, Base says the delay can safely shrink, and it is doing so in stages starting at five days. A faster dual-proof option that settles in about one day already exists but is rarely used because of the cost of generating the required zero-knowledge proof. For bridging providers, a shorter wait frees up capital sooner.
The third change is Reth V2, an updated execution client from Paradigm that Base now runs as its sole node software. Base says the new version cuts node disk usage by roughly half and lets it raise block limits to provide more capacity for developers. The reduced storage and higher throughput are aimed at keeping the network on its longer-term scaling targets.
What participants need to do
Most users and existing smart contracts require no action, Base says. Anyone withdrawing through the canonical bridge will simply wait two fewer days. Node operators must upgrade to the required Base node software before the June 25 activation timestamp or risk falling out of sync with the network. Several exchanges, including Binance, have signaled plans to pause deposits and withdrawals temporarily around the activation window. Developers can begin deploying B20 tokens using the asset or stablecoin variant through Base’s published guides.
Beryl is the second in a planned series of upgrades. Base has scheduled the next one, Cobalt, for September 2026. The team expects Cobalt to add native account abstraction, a protocol-level feature that makes smart accounts standard and allows gas sponsorship and transaction batching without extra contracts, along with further B20 enhancements such as paying network fees in a token other than ether. The four-week gap between Azul and Beryl signals that Base intends to keep shipping quickly as it competes with rival layer-2 networks for stablecoin and tokenized-asset issuance.
Frequently asked questions
What is the B20 token standard?
B20 is Base’s native token standard. It follows the ERC-20 specification so existing wallets and apps work with it, but its logic runs directly in the node software as a Rust precompile rather than as a smart contract, which Base says lowers costs and raises throughput.
When does the Base Beryl upgrade go live?
Beryl activates on Base mainnet on June 25, 2026, at 18:00 UTC. The testnet version has run on Base Sepolia since June 18, 2026. Node operators must update their software before the activation time.
How does Beryl change withdrawals from Base?
Beryl shortens the single-proof withdrawal finalization window from seven days to five for assets moving from Base back to Ethereum. A separate dual-proof path that settles in about one day remains available but is used less often because of its proving cost.
