The field of multi-agent systems is moving towards more efficient and effective workflow automation, with a focus on improving communication efficiency, task performance, and collaboration. Recent developments have introduced innovative architectures, such as hierarchical multi-agent systems and dynamic agent elimination, to optimize workflow generation and execution. These advancements have shown significant improvements in token efficiency, task completion rates, and result quality. Notably, the integration of large language models with specialized tools has enabled the development of more robust and flexible automation systems. Noteworthy papers include: ComfyGPT, which introduces a self-optimizing multi-agent system for comprehensive ComfyUI workflow generation, significantly improving workflow generation precision. AgentDropout, which proposes dynamic agent elimination for token-efficient and high-performance LLM-based multi-agent collaboration, achieving an average reduction of 21.6% in prompt token consumption and 18.4% in completion token consumption. OmniNova, which presents a modular multi-agent automation framework that combines language models with specialized tools, demonstrating outstanding performance in task completion rate, efficiency, and result quality. StarFlow, which explores the use of generative foundation models to automatically generate structured workflows from visual inputs, outperforming large vision-language models on this task. WorkTeam, which proposes a multi-agent NL2Workflow framework that collaboratively enhances the conversion process, significantly increasing the success rate of workflow construction.