Sakana AI Bets on Agent Orchestration Over Frontier Models

Tokyo-based startup Sakana AI is shifting its development focus away from building massive, standalone foundation models toward creating systems that coordinate multiple smaller, specialized AI agents. This approach emphasizes agent orchestration, which allows these autonomous units to collaborate on complex tasks rather than relying on a single, resource-intensive model. By prioritizing modular workflows, the company aims to improve efficiency and reduce the technical barriers associated with deploying large-scale artificial intelligence in enterprise environments.