# Researchers measure a hash-based verifier for Bitcoin bridges

Published 2026-11-16 · Crypto.world Team · Crypto.world

> A preprint measures a garbled, hash-based STARK verifier intended as a post-quantum alternative for Bitcoin bridge disputes. Its example requires 3.93 million virtual bytes across 41 transactions to publish the verifier’s input, while the authors identify further work needed on proof size and verifier setup.

## A hash-based alternative

A preprint measures a garbled, hash-based verifier designed to check STARK proofs in Bitcoin bridge disputes, offering an alternative to verifier designs based on elliptic-curve proofs. The paper, “Toward Efficient Post-Quantum Bitcoin Bridges,” was received by the [Cryptology ePrint Archive](https://eprint.iacr.org/2026/2877) on November 12, 2026, and last revised on November 16 (its fourth revision); the archive labels it a preprint. Its authors are Stephen Duan, Vanhger Pham, Blake Hou, Ethan Zhu and Jingping Yin.

BitVM-style bridge protocols move proof verification off-chain and use a dispute process on Bitcoin. The authors say existing designs garble verifiers for elliptic-curve SNARKs, which a quantum adversary could break. Their proposed alternative uses a hash-based STARK verifier and hash-based one-time keys, avoiding number-theoretic assumptions in the verifier.

The authors compiled a verifier for WHIR-based STARK proofs from Ziren into a Boolean circuit. It uses free-XOR gates for binary-field arithmetic and Blake3 for hashing. A post on Delving Bitcoin, where the research was described, says the design replaces the on-chain verifier in a BitVM2-style construction with a garbled verifier; the operator garbles it once, proves the garbling correct and commits input keys to Bitcoin.

## Measured costs

For an Ethereum mainnet block, number 26,206,588, the example proof is reduced through four recursion levels. The final verifier has 175.4 million non-free gates and reads 520,832 input bits, producing 2.81 gigabytes of garbled tables, according to the paper. The authors say publishing the input with translated Winternitz keys takes 3.93 million virtual bytes across 41 standard transactions.

The Delving Bitcoin post compares that on-chain input publication with a Schnorr adaptor baseline, which takes 1.11 million virtual bytes but is not post-quantum. It also says the garbled verifier is about 15 times smaller off-chain than BitVM3’s Boolean Groth16 verifier. On-chain, however, the proof input is more than a thousand times the roughly 2.4-kilobyte Assert cited for BitVM3.

Completing the Assert in the example takes 43 transactions over at least six blocks: one funding transaction, 41 reveals and a join transaction. The post says the translated-Winternitz approach also requires a 953-megabyte off-chain translation table. A challenged dispute needs about 46 transactions, while an unchallenged claim requires only Kickoff and Take1, according to the same post.

## Unfinished work

The researchers describe the measurements as a trade-off: on-chain publication grows with input size, off-chain storage grows with circuit gate count, and the grinding work used to shorten the input rises exponentially with the grinding level. The ePrint abstract says the team models proof size and proving time to choose settings, while the Delving post details schedule choices that exchange fewer input bits for more expected grinding.

The proposed construction is research, not a reported bridge deployment. The authors list a hash-based proof that the garbling is correct as future work, saying they plan to build and measure it. They also identify the multi-block publication process and a possible batch hash-verification selector as areas for further investigation; the selector would depend on a hypothetical opcode that verifies hashes in batches.

## Sources

1. Original source · Preprint [Cryptology ePrint Archive](https://eprint.iacr.org/2026/2877) eprint.iacr.org

Other pages read

2. [Delving Bitcoin](https://delvingbitcoin.org/t/toward-efficient-post-quantum-bitcoin-bridges/3079) delvingbitcoin.org

## Related stories

- [Blockstream Research reviews lattice-based signatures for Bitcoin](https://crypto.world/news/blockstream-falcon-1024-bitcoin-quantum-candidate)
- [Study finds LLMs often misdiagnose Solana bridge flaws](https://crypto.world/news/study-llms-misdiagnose-solana-bridge-flaws)
- [Preprint: ML-DSA-44 Per-Record Design Costs 8.7× ECDSA Even With Proposed Precompile](https://crypto.world/news/ethereum-ml-dsa-record-signing-calldata-cost)
- [OP Stack EIP-8288 PoC Cuts PQ Data Use From 79 Blobs to 3](https://crypto.world/news/op-stack-eip-8288-pq-signature-aggregation-poc)
- [Preprint proposes one-transaction proofs for Ethereum quantum migration](https://crypto.world/news/researchers-propose-one-transaction-proofs-ethereum-quantum-migration)

## Explore

[More articles](https://crypto.world/news)

---

Canonical: https://crypto.world/news/researchers-measure-post-quantum-bitcoin-bridge-verifier
