Recent advances in text-to-image generators have led to substantial capabilities in image generation. However, the complexity of prompts acts as a bottleneck in the quality of images generated. A particular under-explored facet is the ability of generative models to create high-quality images comprising multiple components given as a prior. In this paper, we propose and validate a metric called Components Inclusion Score (CIS) to evaluate the extent to which a model can correctly generate multiple components. Our results reveal that the evaluated models struggle to incorporate all the visual elements from prompts with multiple components (8.53% drop in CIS per component for all evaluated models). We also identify a significant decline in the quality of the images and context awareness within an image as the number of components increased (15.91% decrease in inception Score and 9.62% increase in Frechet Inception Distance). To remedy this issue, we fine-tuned Stable Diffusion V2 on a custom-created test dataset with multiple components, outperforming its vanilla counterpart. To conclude, these findings reveal a critical limitation in existing text-to-image generators, shedding light on the challenge of generating multiple components within a single image using a complex prompt.