Citations

Use the SEC-bench Pro citation for Pro results and the SEC-bench citation for legacy benchmark-family results.

SEC-bench Pro

Cite this paper when referencing SEC-bench Pro results, datasets, tasks, or harnesses.

@article{lee2026sec,
    author    = {Lee, Hwiwon and Liu, Jiawei and Kim, Dongjun and Zhang, Ziqi and Xia, Chunqiu Steven and Zhang, Lingming},
    journal   = {arXiv preprint arXiv:2605.26548},
    title     = {{SEC-bench Pro: Can Language Models Solve Long-Horizon Software Security Tasks?}},
    year      = {2026}
}

Lee, H., Liu, J., Kim, D., Zhang, Z., Xia, C. S., & Zhang, L. (2026). SEC-bench Pro: Can Language Models Solve Long-Horizon Software Security Tasks? (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2605.26548

Lee, Hwiwon, et al. “SEC-Bench Pro: Can Language Models Solve Long-Horizon Software Security Tasks?” Version 1, arXiv, 2026, https://doi.org/10.48550/ARXIV.2605.26548.

SEC-bench [NeurIPS 2025]

Cite this paper when referencing the original SEC-bench benchmark family and legacy leaderboards.

@inproceedings{lee2025secbench,
    author    = {Hwiwon Lee and Ziqi Zhang and Hanxiao Lu and Lingming Zhang},
    booktitle = {The Thirty-ninth Annual Conference on Neural Information Processing Systems},
    title     = {{SEC-bench: Automated Benchmarking of LLM Agents on Real-World Software Security Tasks}},
    url       = {https://openreview.net/forum?id=QQhQIqons0},
    year      = {2025}
}

Lee, H., Zhang, Z., Lu, H., & Zhang, L. (2025). SEC-bench: Automated Benchmarking of LLM Agents on Real-World Software Security Tasks. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. https://openreview.net/forum?id=QQhQIqons0

Lee, Hwiwon, et al. "SEC-bench: Automated Benchmarking of LLM Agents on Real-World Software Security Tasks." The Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025.