Paper 2026/544

HARE: Compact HQC via Distance-Informed Erasure Decoding

Qicheng Teng, Tsinghua University

Anyu Wang, Tsinghua University

Jun Zhang, Capital Normal University

Bo Pang, ICBC Software Development Center

Chunhuan Zhao, Huawei Technologies

Sihuang Hu, Shandong University

Xiaoyun Wang, Tsinghua University

Abstract

We present HARE, a KEM scheme based on the HQC framework with reduced public key and ciphertext sizes. HARE combines four optimization techniques: (i) a novel distance-informed erasure decoding method for the concatenated code, (ii) ciphertext compression enhanced with covering codes, refining the approach of Bitzer et al. in EUROCRYPT 2026, (iii) unbalanced parameter choices, and (iv) a more accurate error distribution model for decryption failure analysis. Among these, the key innovation is the distance-informed erasure decoding technique, which leverages distance information from the inner code to identify unreliable blocks and treat them as erasures that are subsequently corrected by the outer code. This significantly lowers the decryption failure rate and enables more compact parameter sets. Combining these techniques, HARE achieves substantially improved compactness compared to HQC, reducing the combined size of the public key and ciphertext by 13.7%, 14.0%, and 14.3% at NIST security levels 1, 3, and 5, respectively.

Note: Update the refined DFR analysis based on hypergeometric distribution

@misc{cryptoeprint:2026/544,
      author = {Tianrui Wang and Qicheng Teng and Anyu Wang and Jun Zhang and Bo Pang and Chunhuan Zhao and Sihuang Hu and Xiaoyun Wang},
      title = {{HARE}: Compact {HQC} via Distance-Informed Erasure Decoding},
      howpublished = {Cryptology {ePrint} Archive, Paper 2026/544},
      year = {2026},
      url = {https://eprint.iacr.org/2026/544}
}