Muse Spark Open: Meta's 816B open-weight model for coding and agents
What is it?
Meta took the Muse Spark model it released in August, added more safety post-training, and published the weights. It's an MoE (mixture-of-experts: only some expert blocks fire for each input) with 816B total parameters and about 42B active per token. It accepts text, images, video and audio, and has a 1M-token context window. Weights are on Hugging Face as meta-models/Muse-Spark-Open-816B-A42B in BF16 and NVFP4, and SGLang, vLLM, TensorRT-LLM, MLX, llama.cpp and Unsloth support it from day one. With llama.cpp:
llama serve -hf ggml-org/Muse-Spark-Open-816B-A42B-GGUF:Q4_K_SWhy does it matter?
In Meta's own table it scores 89.2 on MCP-AtlasVendor claim and 76.4 on Terminal-Bench 2.1Vendor claim. It beats Kimi K3 and GLM 5.3 on tool use but trails Kimi K3 (85.0) on Terminal-Bench, so try it first for tool-heavy business agents rather than pure coding. It runs on two M5 Ultra Mac Studios with 512GB each, or in roughly 300GB of memory with a 2-bit GGUF, so internal agents that can't send data off-site finally have a serious open option.
Maturity
GAWeights are downloadable and hosted APIs are live on Fireworks, Together, OpenRouter and others, so this is a full release. The license is Meta's own Muse Model License, not Apache or similar, so check its terms before you serve it as a product.
How to track it
- Independent scores: see whether Muse Spark Open shows up on artificialanalysis.ai and whether its ranking matches Meta's table.
- License: read the Muse Model License terms on the Hugging Face model card for commercial-use conditions.
- Real-world speed: watch the Hugging Face discussion tab for Mac Studio and workstation numbers from Unsloth and llama.cpp users.