The field of multi-agent systems is moving towards developing more efficient and scalable methods for communication and coordination among agents. Researchers are focusing on addressing the challenges of insufficient information flow and limited information processing capacity in dynamic collaborative systems. New frameworks and algorithms are being proposed to optimize communication, reduce token consumption, and improve overall system performance. Notable advancements include the development of information flow structures, group sequence policy optimization, and multi-agent system process reward models. These innovations have shown promising results in improving the efficiency and robustness of multi-agent systems. Noteworthy papers include: IFS, which proposes an information flow structure to improve information flow and processing capacity in multi-agent ad hoc systems. Agent-GSPO, which introduces a framework for communication-efficient multi-agent systems via group sequence policy optimization. MASPRM, which presents a multi-agent system process reward model to guide inference-time search and selectively spend compute to improve quality. MARs, which proposes a multi-party agent relation sampling method for multi-party ad hoc teamwork. SupervisorAgent, which introduces a lightweight and modular framework for runtime, adaptive supervision to enhance robustness and efficiency in multi-agent systems.