How Sakana trained a 7B model to orchestrate GPT-5, Claude Sonnet 4 and Gemini 2.5 Pro

Sakana AI has developed the 'RL Conductor,' a 7B model that utilizes reinforcement learning to effectively orchestrate multiple large language models (LLMs) such as GPT-5 and Claude Sonnet 4. This innovative approach allows for automated coordination of tasks among LLMs, resulting in superior performance on complex reasoning and coding benchmarks while reducing costs and API calls compared to traditional methods. The RL Conductor represents a significant advancement in the efficiency of AI model orchestration.