None
None
Long Story Video Skill
Turn your idea into 10s–10min video—AI writes scripts, prompts, and generates footage automatically.
Ads Video Skill
Generate professional ads and sales videos—AI auto-generates scripts, prompts, and footage.
3D Science Explainer Video Skill
Convert scientific concepts into stunning 3D explain animations
Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
Muse Spark Open
Explore Muse Spark Open: an 816B open-weight multimodal model with a 1M-token window that runs agentic video edits right on your Mac.
All Tools
Discover our comprehensive AI-powered animation toolkit
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.
Seedance 2.0
The Future of AI Video Is Here.
Veo3.1
Create Stunning Videos with Veo3.1
FLUX 3 Video Generator
Kling 3.0
Next-Gen AI Video Generator
Grok Video Generator
Create Videos from Text or Images with AI
MiniMax H3 video generator
MiniMax H3 AI Video Generator
Inside the Muse Spark Open Release
Shipped as fully open weights, Muse Spark Open is a natively multimodal engine built for coding, autonomous agents and long-horizon jobs that you host on hardware you control.
- Weights You Download and OwnThe strongest open-weight drop yet from a US lab: 816B total parameters, 42B active, a million-token window, and multimodality baked in from the start rather than layered on later.
- Frontier Power on Your DeskClass-leading models should not live only behind a datacenter wall. Muse Spark Open boots on a pair of Apple M5 Ultra Mac Studios paired with the newest MLX runtime.
- Agentic Work, Proven LocallyRunning through Muse Code, it shipped genuine iOS fixes and features into Ice Cubes, then drove DaVinci Resolve start to finish and auto-cropped the strongest frames to vertical.
Three Ways to Deploy Muse Spark Open
Pick the path that fits your setup — local Apple silicon, an inference engine you already trust, or a managed endpoint in the cloud.
Core Strengths of Muse Spark Open
A million-token window, built-in multimodality, desktop-grade memory bandwidth and downloadable weights — the ingredients that make agentic video production possible.
One Million Token Context
The 1M-token window keeps long agentic sessions and multi-step reasoning chains anchored to earlier decisions instead of losing the thread halfway through.
Agentic Coding That Lasts
Engineered for software engineering and self-directed loops, Muse Spark Open sustains long-horizon jobs and drives end-to-end control of tools such as DaVinci Resolve.
Multimodal by Design
Images, footage and text are consumed natively rather than through add-on adapters — precisely why frame-by-frame review of raw video actually works.
Two Macs, One Supercomputer
Unified memory lets every M5 Ultra GPU reach up to 512 GB, so a pair of Mac Studios pools roughly a terabyte and holds the entire model as a single machine.
1.2 TB/s of Memory Bandwidth
That throughput feeds the small active slice computed for each token, which is what keeps Muse Spark Open feeling snappy on a desktop workstation.
Safety Tuned Before Release
Extra safety post-training was applied for the open release: the model was scored on the Meta Superintelligence Scaling Framework and passed its bar across every measured risk category.
Muse Spark Open: Your Questions Answered
What people ask most about hardware requirements, licensing, downloads and deployment for Muse Spark Open.
Which setup do I need to run Muse Spark Open locally?
Two Apple M5 Ultra Mac Studios running the newest MLX build. Each one offers up to 512 GB of unified memory, giving roughly a terabyte combined for the full checkpoint.
How big is the model, and how much context can it hold?
816B total parameters with 42B active, a 1M-token context window and native multimodality — aimed squarely at coding, agentic loops and long-horizon work.
Where do I get the weights, and is a hosted API available?
Download them via Hugging Face, Amazon SageMaker AI, Google Cloud Model Garden or Microsoft Foundry Models; managed endpoints run on Baseten, Fireworks AI, Groq, Modal, Nebius, OpenRouter and Together AI.
Which inference engines and chips are supported?
Take your pick from MLX, llama.cpp, Unsloth, SGLang, vLLM, TensorRT-LLM or Transformers — tuned across NVIDIA, AMD, Intel and Apple hardware, with PyTorch underneath whenever you want to go deeper.
What license covers the released weights?
The Muse Model License applies. Its limited commercial terms still leave broad room for developers to run Muse Spark Open without paying a fee.
How was safety evaluated before launch?
It was scored against the Meta Superintelligence Scaling Framework covering the most severe risks, received additional safety post-training for open weights, and cleared the release thresholds — the model card holds the detail.
Start Building with Muse Spark Open
Pull the checkpoint from Hugging Face, query it through hosted inference partners, and share what you make with the community on the Hugging Face Forums.