TechAILatest.newsTrackers

Open-weight multimodal language model

Muse Spark Open

Muse Spark Open is an open-weight mixture-of-experts model from Meta, released on 15 October 2026 with 816B total and 42B active parameters and a 1M-token context window. It accepts text, image, video and audio input, outputs text, and is licensed under the Muse Model License rather than a standard open-source license.

Last reviewed 28 Oct 2026 · by Meta

Facts at a glance

DeveloperMeta (Meta Superintelligence Labs)
Released15 October 2026 source
Parameters816B total, 42B active per token (256 experts, 8 active) source
Context window1M tokens source
ModalitiesText, image, video and audio input; text output
LicenseMuse Model License (attribution required; separate agreement for Model as a Service above $20M revenue)
WeightsBF16 and NVFP4 on Hugging Face (gated access) source
HardwareAt least 700GB of VRAM for NVFP4 and 2TB for BF16, per the model card source

What Muse Spark Open is used for

Meta describes Muse Spark Open as a natively multimodal model built for coding and agentic work, such as coding agents and customer service agents. The weights run in SGLang, vLLM, TensorRT-LLM, Hugging Face Transformers, Apple MLX, llama.cpp and Unsloth, and hosted inference is offered by providers including Baseten, Fireworks AI, Groq, Modal, Nebius, OpenRouter and Together AI. Meta says the model can run on two Mac Studio machines with M5 Ultra chips using MLX.

Questions people ask

Is Muse Spark Open open source?
Meta publishes the weights, but under its own Muse Model License. The license allows use, modification, distillation and sale, requires a "Built with Muse Spark Open" notice, and requires a separate agreement for Model as a Service businesses with more than $20 million in annual revenue.
How large is Muse Spark Open?
816B total parameters with 42B active per token, according to Meta's model card, plus a 1.8B-parameter perception encoder.