Llama 4 Scout (8B - 70B)

A high-context model optimized for deep research. It supports a 1M+ token context window, allowing for the ingestion of entire document libraries for institutional-scale analysis.