MODEL REFERENCE / meta-models
meta-models/Muse-Spark-Open-816B-A42B
Sparse MoE reasoning model with 816B total / 42B active parameters and a 1M-token context, built for long-horizon agentic coding. Native image and video perception, channel-scoped reasoning and tool calls via the `musespark1x` parsers, and an optional DSpark draft model for speculative decoding.
MXFP4 and BF16 on one 8-GPU node: TP + expert parallel with DSpark for interactive serving, TP + expert parallel alone for peak MXFP4 throughput
View on HuggingFace
View on ModelScope02 / Configure
Configure serving
Choose engine, hardware and deployment options. The command updates as you go.
03 / Understand