this is the part people sleep on. meta just gave away an 816b parameter model, 42b active, 1m token context, multimodal, built for coding agents. the weights are free. running them is not cheap. card breaks it down.
Today we’re releasing Muse Spark Open, the most capable US open-weight model to date.
Built for coding and agentic work, it excels at long-horizon tasks, complex software engineering, and multi-step reasoning.
Key Model Specs:
→ 816B total / 42B active parameters
→ 1M token context window
→ Natively multimodal across text, image, video, audio