zkID Benchmarks v1.1 (Jun 2025)

[v1.1]

🐛 Fixed

• Restored the RAM measurement in Binius benchmarks

Full Changelog: zkid-benchmarks-v1...zkid-benchmarks-v1.1

zkID Benchmarks v1 (Jun 2025)

TL;DR: Binius and Ligero emerge as the top candidates for mobile SHA-256 ZK-proving under transparent, post-quantum constraints. This work is helpful for wallet developers, standards authors, and ZK library maintainers in the digital ID space who require a post-quantum-secure proof scheme with a transparent setup that fits within tight mobile CPU, RAM, and bandwidth budgets. We compare seven schemes on a SHA-256 circuit with 2 kB input, measuring prover time, RAM usage, proof size, and preprocessing overhead. Our results show that:

• Binius offers ≈5 s proving, sub-50 MB RAM, and small proofs.
• Ligero is a runner-up with ≈30-95 s proving and a modest RAM/VRAM footprint.

No single scheme is perfect: Binius excels at bitwise operations, while Ligero achieves performance by utilizing WebGPU. Developers should choose based on their specific mobile-context trade-offs. Our future work will build upon these benchmarks, so feel free to reach out to the Client-Side Proving team if you have a particular use case or would like to collaborate.

More info: https://pse.dev/blog/efficient-client-side-proving-for-zkid