Multi-Agent with Small Models

Use multi-agent architectures around smaller open-source GPT models to orchestrate systems-level workflows and compensate for individual model limitations, improving overall performance.