Scale Agentic Coding Workflows with Meta’s Muse Spark Open on NVIDIA GB300 NVL72
Meta launches Muse Spark Open, an 816B MoE coding model with 1M token context window
Meta released Muse Spark Open, an 816B mixture-of-experts model with only 42B active parameters and a 1M token context window, purpose-built for agentic coding workflows on NVIDIA GB300 NVL72 Blackwell Ultra hardware. The model targets higher first-attempt accuracy, reliable tool calling, and whole-repository comprehension. As an open model at this scale optimized specifically for agentic use cases, it directly challenges closed coding models and raises the bar for enterprise AI development tooling.