@article{cudaherculesbenchmarkinghardwareawareexp, title = {CUDAHercules: Benchmarking Hardware-Aware Expert-level CUDA Optimization for LLMs}, author = {Shiyang Li and Zijian Zhang and Guangyan Sun and Yuebo Luo and Winson Chen and Yanzhi Wang and Mingyi Hong and Caiwen Ding}, year = {2026}, eprint = {2605.08467}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2605.08467}, }