leanXMSS Test Reaches 1,191.353 Signatures per Second on Post-Quantum Ethereum Project
Key Facts
The leanVM project page reports a test that aggregated 1,400 signatures using the leanXMSS system at a stated rate of 1,191.353 signatures per second, an early technical indicator in Ethereum’s research to address quantum threats. The repository says proof generation in the test took 1.175 seconds and that the measurement was taken on a MacBook Pro with an M4 Max processor, without using the Metal GPU. But the same page says formal verification of leanVM is still in progress and the system is not yet production-ready. The results therefore describe the performance of a specific test and do not mean the technology has become part of the live Ethereum network. The page presents the project as a hash-based zero-knowledge proof virtual machine aimed at post-quantum Ethereum.
The project page distinguishes leanVM’s performance from the signature aggregation measurement: the published figure concerns aggregation of leanXMSS signatures, not data processing or network speed. It also cautions that the parameter set used in the leanXMSS test is not final, meaning performance results could change as the design evolves. The repository says leanXMSS parameters are documented in a separate file and have a security proof in the ROM model written in Lean 4. The existence of that proof or the test result does not make the system production-ready. According to the project description, leanVM verification remains in progress.
Documentation for leanerVM, which is working to formalize leanVM in Lean 4, says some proof layers are not yet complete. Specifically, the documentation says bus theorems and two theorems in the T1 layer remain to be completed, and a claim related to the proof system has not yet been proved. The project aims to build soundness and completeness proofs from the virtual machine implementation through the proof system and recursion, as well as verify Rust application behavior. Current verification starts by comparing the implementation with specified Rust versions, with the possibility of later expanding to formal verification of the relevant code. This status supports describing the work as ongoing research and development, not a completed network upgrade.
The work sits within a broader Ethereum Foundation push toward hash-based cryptography, amid what Justin Drake described as the rapid spread of SNARKs and an imminent quantum threat. In the foundation’s vision for upgrading network layers, the consensus layer would use hash-based aggregated signatures to replace BLS signatures. A roadmap published in 2026 sets a goal of quantum resistance across the execution, consensus and data layers by December 2029, with progress to be reassessed in January 2027. The foundation’s J* milestone includes a quantum-resistant pulse in the consensus layer, leanDA sampling in the data layer and leanSPHINCS transactions in the execution layer. Full quantum resistance across the layers remains a future goal in the plan.
The foundation says a 12-month upgrade cadence starting with the Glamsterdam upgrade could enable it to reach the minimum quantum protection milestone by December 2029. It describes the milestone as temporary protection that would allow Ethereum to continue operating on a day when quantum computing makes a decisive advance, with fewer guarantees than full protection. This phased path runs alongside development of technologies such as leanVM and leanXMSS signature aggregation, without implying that these components are ready or approved for the network. Drake is scheduled to take part in a session on Ethereum’s post-quantum roadmap at Devcon on 4 November 2026, from 12:30 to 12:55. The session will provide a near-term opportunity to follow details of the announced path, while implementation and readiness dates remain tied to progress on the work.
In markets, Ethereum closed at $2,288.74 against the dollar on 3 November 2026. The close came in the session before Drake’s scheduled Devcon session, while the available leanVM data focus on technical progress and a specific performance test. These figures alone do not show that a security upgrade has gone live or that the development has had a direct market impact. A distinction remains between the foundation’s goal of quantum resistance across network layers and leanVM’s status, which the project page describes as not production-ready. The market close provides a dated figure for the asset’s performance in that session, without linking it causally to the development work.