Ethereum's upcoming Glamsterdam upgrade has passed a critical rehearsal phase, with Nethermind's execution software completing all 2,302 performance tests ahead of a planned Oct. 6 deployment on the Sepolia public test network.
Nethermind reported Wednesday that its client processed 570.7 billion gas in three minutes and 15 seconds — averaging roughly 2.9 billion gas per second — during benchmarking of one of Glamsterdam's most significant changes. The tests evaluated block-level access lists, a mechanism that provides advance notice to Ethereum nodes about which accounts and stored data a given block will use before processing begins.
That advance information allows nodes to fetch data and verify unrelated transactions in parallel, a departure from Ethereum's current mostly sequential processing model that constrains how much activity can safely fit into each block. The benchmark measured Nethermind's execution software specifically, not full network speed.
Glamsterdam raised its testnet block gas limit from 60 million to 200 million approximately an hour after going live. Gas measures the computational work required by transactions, and a higher ceiling creates additional room for payments, token swaps and other on-chain activity within each block. However, the 200 million figure remains a test setting and is not yet planned for Ethereum's mainnet.
A larger block gas limit would enable Ethereum to absorb more transaction volume before users begin outbidding each other for block space, potentially softening fee spikes during periods of heavy trading or when popular token launches congest the network. Larger blocks are also more computationally demanding to verify, raising concerns that smaller node operators could struggle to keep pace. Glamsterdam aims to deliver additional capacity without making the chain prohibitively expensive to run.
The upgrade restructures the relationship between specialized builders, who package transactions, and validators, who check resulting blocks and secure the chain. Under Glamsterdam, that handoff and accompanying payments would move inside Ethereum's own protocol rules, reducing reliance on third-party relay services. The change also extends the window for distributing block data across the network from about two seconds to nine seconds, giving validators more time to receive and validate the larger blocks the upgrade is designed to support.
Devnet-11, which served as the most recent rehearsal environment, was structured as a controlled test without deliberate attack scenarios. The core test network — where the 200 million gas limit was observed — continues confirming blocks under the new block-building rules.
The proposed Oct. 6 rollout on Sepolia represents another milestone on the path toward a mainnet deployment, though no firm date has been set for that release.












