With the widespread application of deep neural networks, generative models, especially diffusion models and flow matching models, for protein design have experienced explosive growth. However, there remains a lack of comprehensive evaluation frameworks to systematically assess the performance of these models. This study addresses this gap by focusing on the task of designing unconditional protein structures, benchmarking seven state-of-the-art (SOTA) models in four distinct dimensions: structural validity, diversity, novelty, and computational efficiency. This work provides standardized metrics and baseline benchmarks to guide future research and innovation in protein design.