Meta's open-weight 816B-parameter mixture-of-experts model (42B active, 1M context) in the Muse model family.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Muse Spark Open is an open-weight language model from Meta, announced on research.meta.ai as "the most capable US open-weight model to date". According to a YouTube creator summary, it is a mixture-of-experts model with 816B total parameters, 42B active per token, and a 1M-token context window. It belongs to Meta's Muse family, which also includes Muse Spark, Muse Glimmer, Muse Ember and Muse Wonder.
The same creator noted that on an independent intelligence index it ranks below other open models such as GLM 5.3, and that it costs more to run. The license is described as restrictive, requiring attribution and registration. Exact terms, benchmark figures and download locations were not in the provided page text, so these details should be checked against Meta's release post.
It is aimed at developers and organisations that want to self-host or fine-tune a large model, rather than only using Meta's hosted Muse Spark products.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Open-weights 816B MoE model (42B active, 1M context) claimed to be the most capable US open-weight model. It ranks below other open models like GLM 5.3 on the intelligence index, costs more, and has a restrictive license with required attribution and registration.
Open-weight release of Muse Spark, 816B parameters with 1M context. Meta calls it the most capable US open-weight model, and it was optimized with Apple to run on an M5 Ultra Mac Studio. It tops most open-weight comparisons except Kimi on the coding benchmark.
Closest neighbours by meaning, not just by category.
Model · Meta
Meta's open-weights 30B dense model for local agents, tuned for tool use and long tasks, Apache 2.0 licensed.
Model · Meta
Meta's Muse Spark 1.3 language model, described by creators as near-frontier on benchmarks.
Model · Meta
Meta's proprietary large language model, cited by creators as a top US-made model on the Artificial Analysis index.
Model · Meta
Closed-source Meta model for coding and agentic workflows with 1M-token context and multimodal input, offered via API.
Model
35B Mixture-of-Experts open-weights model from InternScience for long-horizon agentic search, engineering and tool use.
Model
Compact 1.7B open-weights text LLM for on-device use, with adjustable thinking levels and English/Chinese support.
