Muse Spark Open: what the license allows, and what the maker charges
thecontinuitylayer.com/open-models/muse-spark-openMeta's open-weight model from Meta Superintelligence Labs, released 15 October 2026: 816B total and 42B active parameters with a 1M-token context window. It is Meta's newer open release than the Llama 4 models of April 2025.
- 1Partly
Is Muse Spark Open under an open license?
Partly, by Meta's own page.
The weights are under the Muse Model License, a custom license: it grants use, modification, fine-tuning and distribution, requires "Built with Muse Spark Open" in the user interface, incorporates a Usage Policy, and asks a model-as-a-service business with over $20 million in gross revenue over 12 months to sign a separate agreement with Meta.
Yes means: The weights are under an OSI-approved license such as Apache-2.0 or MIT. Partly means the maker's own license with conditions, such as a user cap or an acceptable-use clause; no means research or non-commercial use only.
- 2Partly
Can you download Muse Spark Open's weights?
Partly, by Meta's own page.
BF16 and NVFP4 weights are on Hugging Face under meta-models, gated: you log in, accept the conditions and submit a form with your legal name, date of birth and organization before access.
Yes means: The maker's page links the weights for download with nothing more than a click-through license. Partly means an access form, an approval step, or only some sizes released.
- 3Yes
Does the maker state Muse Spark Open's context window?
Yes, by Meta's own page.
The card states a context window of 1M tokens, "1M at inference."
Yes means: A token count for the context window is on the maker's page or model card.
- 4Yes
Does the maker say what it takes to run Muse Spark Open on your own machine?
Yes, by Meta's own page.
816B total and 42B active parameters; the card says NVFP4 needs a GPU with at least 700GB of VRAM and BF16 needs 2TB, and that the model is optimized to run on two Mac Studios with M5 Ultra using MLX.
Yes means: The page states the parameter count and the memory or hardware it needs. Partly means the parameter count alone.
- 5No
Does the maker publish a price for its hosted Muse Spark Open?
No, by Meta's own page.
Meta's Model API page says of Muse Spark Open: "You don't call it over Model API"; you self-host it or use a hosted inference provider. Meta's per-token prices cover the closed Muse Spark 1.1, 1.2 and 1.3.
Yes means: A per-token price on the maker's own pricing page. Partly means a hosted version exists but the price sits behind sign-in or a sales call.
- 6Not stated
Does the maker train on your prompts when you use its hosted Muse Spark Open?
Meta's pages do not say.
Meta does not host this model, so there is no hosted default to grade. For the closed Muse Spark on Meta's Model API, Standard mode does not use your content to train Meta models and Contributor mode may, at a lower price.
Yes means: Here no is the good answer: the maker's page says what you send its API is not used for training. Yes means it is unless you opt out; partly means it depends on the plan or the page does not give the default.
- 7Partly
Can you fine-tune Muse Spark Open and keep the result?
Partly, by Meta's own page.
The license grants "the rights to use, copy, modify (including fine-tune), create derivative works, distill, merge, publish, distribute, sublicense, and/or sell copies of the Software", with conditions: the "Built with Muse Spark Open" display, the Usage Policy, and the $20 million model-as-a-service clause.
Yes means: The license allows fine-tuning and the fine-tuned weights are yours to keep and share. Partly means conditions follow the derivative, or a hosted fine-tune you cannot download.
License as stated: Muse Model License, version Muse Spark Open, release date 15 October 2026.. Every grade above is read from Meta's own page on the day shown. If a page has changed, tell us.