This week we cover zero-knowledge proofs for LLM inference, a Lean-certified STARK verifier in OpenVM, and Ethereum block-proving benchmarks from Pico and Ziren. The papers explore bit packing and ring switching over binary fields, threshold anonymous credentials, and the conditions for preserving Fiat–Shamir extraction when instantiating the random oracle.

• Roy Rinberg, Alex Allain, Alessandro Chiesa, Miranda Christ, Rahul Dalal, Giacomo Fenzi, Ben Mayer, Guy Weissenberg

• An open-source system for proving LLM inference while keeping model weights private.

• https://www.anthropic.com/research/provable-inference

• https://www.anthropic.com/document/flashprove-technical-report

• Adds a Lean-certified STARK verifier to the OpenVM SDK and CLI.

• https://github.com/openvm-org/openvm/releases/tag/v2.1.0

• A Pico zkVM release with a reported 128-bit security bound and an open-source GPU prover.

• https://blog.brevis.network/2026/11/05/pico-3-1-128-bit-provable-security-still-real-time-on-4-gpus-now-fully-open-source/

• A MIPS32 zkVM release with a proving core based on jagged sumcheck and WHIR.

• https://www.zkm.io/blog/ziren-2-0

• https://hoseinhadipour.com/blog/lean-for-cryptanalysts/

• Matteo Frigo, abhi shelat

• Uses the field trace to check linear claims about witness bits packed into binary extension-field elements.

• https://eprint.iacr.org/2026/2770

• Sebastian Faust, Jonas Hofmann, Philipp-Florens Lehwalder, Hussien Othmann

• Builds threshold anonymous credentials from Boneh–Boyen signatures without random oracles.

• https://eprint.iacr.org/2026/2762

• Andrei Kochergin

• Gives a zero-knowledge version of Diamond and Posen's ring-switching technique.

• https://eprint.iacr.org/2026/2754

• Ashrujit Ghoshal, Sikhar Patranabis

• Studies when Fiat–Shamir knowledge extraction is preserved as a structured hash replaces the random oracle.

• https://eprint.iacr.org/2026/2708

• Tianyu Zhang, Nikita Borisov, Yupeng Zhang

• A zkSNARK for native fixed-width integer arithmetic using reverse multiplication-friendly embeddings.

• https://eprint.iacr.org/2026/2672

• Ignacio Cascudo, Carmit Hazay, Muthu Venkitasubramaniam

• Sharpens Ligero's soundness analysis and introduces a variant that supports higher Reed–Solomon code rates.

• https://eprint.iacr.org/2026/2681

• Kyeongtae Lee, Jihye Kim, Hyunok Oh

• A Shockwave-based prover that uses sequential storage to keep RAM usage low during indexing and proving.

• https://eprint.iacr.org/2026/2689

• Zhi Lu, Songfeng Lu, Jiaheng Zhang

• Uses prover randomness to let a public proof also certify knowledge of a hidden witness to a designated verifier.

• https://eprint.iacr.org/2026/2706

• Hamidreza Khoshakhlagh, Anders Kallesøe, Anders Dalskov

• Uses incremental proofs and an external prover to let auditors verify state updates and committee rotations in long-running MPC.

• https://eprint.iacr.org/2026/2721

If you're interested in our ZK Insights or have ideas for similar content, we highly encourage everyone to head over to our GitHub repo and submit a Pull Request. Join forces with like-minded ZKPunks to co-create!

✨ Web collection version: https://insights.zkpunk.pro/
✨ GitHub repo link: https://github.com/ZKPunk-Org/zk-insights