aiqbench takeMachine Learning

The release adds day-0 vLLM support for Muse Spark Open, an 816B-parameter open-weight mixture-of-experts model with 42B active parameters and a 1M-token context window. For practitioners, the useful details are the NVIDIA, AMD, and initial TPU serving paths, the DSpark speculative-decoding draft model, and the fact that the code lands in nightly wheels only after PR #61879 is merged, likely about a week out. Benchmark claims such as "most capable US open-weights model" are the vendors' own and should be checked against independent evaluations.

Sources coveringvLLM Blog, SiliconANGLE, Latent Space

Day 0 support for Muse Spark Open on vLLM

vLLM BlogOctober 15, 20266 min read

Day-0 vLLM support for Muse Spark Open: 816B MoE, 1M context, multimodal inputs, DSpark speculative decoding, NVIDIA and AMD serving, initial TPU support, and LoRA RL.

We're excited to have worked with Meta Superintelligence Labs to bring Muse Spark Open to vLLM on day 0. Muse Spark Open is an open-weight sparse mixture-of-experts (MoE) model, built for long-horizon agentic coding workflows. It is the most capable US open-weights model to date and it is the model Muse Glimmer 30B was distilled from. Architecture: 816B total parameters, 42B active per token with a 1M context window. Quantization: Meta released the model in both BF16 and NVFP4. MXFP4 weights were released by AMD. Multimodal: Native image and video perception through the same Perception Encoder as Muse Glimmer, with additional audio support. Parallelism: Tensor Parallelism and Pipeline Parallelism on both Nvidia and AMD. EP, TEP, DEP and PD disaggregation on Nvidia. Initial support on TPU…

Excerpt shown under the publisher's RSS terms · 126 words · full text and figures are on the publisher's site

Read at vLLM Blog ↗Ask about this article