Our paper, MobileKernelBench: Can LLMs Write Efficient Kernels for Mobile Devices?, is now available on arXiv.