Reth v2.7.0 prepares Sepolia for the Glamsterdam hardfork and brings fixes to state persistence, forkchoice, RPC, and networking. Action required Sepolia operators must upgrade before 6 October 2026 at 13:53:36 UTC. This release schedules Amsterdam activation on Sepolia testnet and tests the fork ID and eth_config transition. (Reth change, official schedule) Highlights Hardfork and storage readine…
Reth v2.6.0 expands Amsterdam and Hegotá protocol support, installs the REST-SSZ Engine API by default, advances EIP-8189 snap-sync infrastructure, and moves partial persistence further into production paths. The release also improves performance across block processing, historical-state RPC, pruning, transaction-pool updates, and peer discovery. On a 300M-gas BAL workload, storage-trie job parall…
Maintenance Release Recovers ExEx WALs with non-contiguous file IDs, preventing startup failures after finalization leaves gaps. Allows networks with dynamically scheduled forks to update their advertised fork filter without restarting. Improves RocksDB open-file limit handling and includes rejected block hashes in stage error logs. Update Priority User Class Priority Payload Builders High Non-Pay…
Maintenance Release Finalizes pending payload builds when resolved early, preserving accumulated work instead of discarding it. Correctly handles whole-storage trie wipes during batched persistence. Update Priority If you are running on a low-core machine (<5 cores), the update is mandatory as it fixes a regression. User Class Priority Payload Builders Medium Non-Payload Builders Medium All Change…
Reth v2.5.0 improves execution performance across the board. Release regression testing against v2.4.1 measured 5-10% lower mean block-processing latency, especially in long-tail workloads. Performance Added an opt-in sender recovery cache shared across transaction paths, reducing mean mainnet block-processing latency by 3%. Enable it with --engine.sender-recovery-cache (#26494). Added an opt-in t…
Latest Release This patch release is for alloy dependency updates. Update Priority This table provides priorities for which classes of users should update particular components. User Class Priority Payload Builders High Non-Payload Builders High See Update Priorities for more information about this table. All Changes chore: remove release formatting noise chore(deps): bump alloy core 1.6.1 (#26419…
Performance Reth v2.4.0 reduces mean newPayload latency by 6.2% compared with v2.3.0. Experimental revmc JIT EVM support is added and can be activated with --jit.enabled flag. Networking and transaction propagation get a large set of throughput-oriented improvements: faster pooled transaction recovery (#26304), cached full transaction propagation encoding (#26012), and multiple ECIES / P2PStream a…
Reth v2.3.0 includes several payload, engine, trie, and transaction-pool performance improvements, including parallel execution work, BAL prewarming, faster proof and trie paths, cheaper transaction-pool insertion, and reduced payload-building overhead. In our benchmark, throughput improved from ~1.4 to ~1.5 Ggas/s, an 8.1% increase. It also fixes important correctness issues in eth_simulate, proo…
chore: relax limits
This release lays the groundwork for Amsterdam (EIP-7928 Block Access Lists), introducing a BAL store abstraction, P2P wire support for fetching BALs alongside full blocks, payload-builder integration, and parallel/batched BAL execution paths. It also enables Discv5 by default, adds snap/2 wire helpers, and bounds the memory footprint of the P2P transaction manager and network channels. Numerous p…
Importantalloy-evm bump fixes an issue where gas refunds issued by custom precompiles were not accounted for. If your network was setting gas_refunded field to a non-zero value for any precompile, its behaviour might change. Storage v2 Migration This release includes a reth db migrate-v2 command that can be used to migrate from Storage v1 to Storage v2 without a full resync. For Ethereum Mainnet,…
Read our Blog Post Reth 2.0 is a major release that makes Storage V2 the default for all new nodes, introduces the SparseTrieCacheTask for dramatically faster state root computation, and delivers across-the-board performance improvements to the engine, RPC, and P2P layers. This release also removes the deprecated reth-primitives crate and removes RocksDB feature gates (RocksDB is now always availa…
Latest Release ImportantThis release replaces v1.11.x, and users must upgrade to v1.11.3 fix(trie): panic in sparse trie proof workers Update Steps Nodes can be restarted using the new binary with no required update steps. Errors during payload validation may continue to be seen, but will not stall the node. To fully fix errors such as those described in issue 22973, shut down the node and run: re…
chore: release 1.11.3
Latest Release ImportantThis release replaces v1.11.1, and users must upgrade to v1.11.2 fix(engine): reset execution cache hash on clear (#22895) Backwards Compatibility ImportantIn case you need to downgrade after running v1.11.x, please downgrade to the latest v1.10.x release. If you need to downgrade further below v1.10.0, please downgrade to v1.9.4. Update Priority This table provides priorit…
Latest Release ImportantThis release replaces v1.11.0, and users must upgrade to v1.11.1 fix: properly reveal trie nodes (#22415) fix: overlay preparation on tokio (#22492) Backwards Compatibility ImportantIn case you need to downgrade after running v1.11.x, please downgrade to the latest v1.10.x release. If you need to downgrade further below v1.10.0, please downgrade to v1.9.4. Update Priority T…
Reth v1.11.0 delivers massive state root performance gains, benchmarks show a ~25% reduction in mean newPayload latency (42.9ms → 32.4ms), ~27% improvement at P90 (72.4ms → 53.1ms), and +33% throughput (700M → 1G gas/s), through the new sparse trie as cache. Up-to-date performance metrics can be seen on the ethpandaops Labs dashboard. v1.11.0 also introduces an experimental hot/cold storage archit…