IBM's modular Granite model, released 14 October 2026 in 8B and 30B sizes: one endpoint for chat, scored multiple-choice decisions, safety checks and English speech-to-text, under Apache 2.0.
Muse Spark Open
New· 4Meta's open-weights Muse model, released 15 October 2026: 816B parameters with 42B active, a 1M-token context, text, image, video and audio input, and a licence that is free below a $20M service threshold.
- United States
- Artificial Intelligence - AI
- Machine Learning
- Natural Language Processing
- +1 more
TrustList summary
Muse Spark Open is the open-weights version of Meta's Muse Spark model, released by Meta Superintelligence Labs on 15 October 2026. Meta calls it the most capable US open-weight model to date and aims it at coding and agentic work: long-horizon tasks, complex software engineering and multi-step reasoning. It is a mixture-of-experts model with 816 billion parameters, of which 42 billion are active for each token, and it holds a context of up to one million tokens.
What it does
The model reads text, images, video and audio and returns text; it does not generate images, sound or video. Meta ships the weights in BF16 and in an NVFP4 build that quantises only the experts. The model card asks for at least 700 GB of accelerator memory for NVFP4 and 2 TB for BF16. Meta's headline deployment is a desktop one: two Apple Mac Studios with the M5 Ultra chip, each with up to 512 GB of unified memory, run the model through Apple's MLX framework, and Meta describes on-device tests in which its Muse Code agent made changes to an open-source iOS app and edited video in DaVinci Resolve. In the data centre it runs on SGLang, vLLM, TensorRT-LLM, Hugging Face Transformers, llama.cpp and Unsloth, with optimisations for AMD, Apple, Intel and NVIDIA hardware.
Who it is for
Meta says hosted access is available on day one through Baseten, Fireworks AI, Groq, Modal, Nebius, OpenRouter and Together AI, and the weights can be downloaded from Amazon SageMaker AI, Google Cloud Model Garden, Hugging Face and Microsoft Foundry Models. The Muse Model License allows use, fine-tuning, distillation and redistribution, with two conditions a business should read: products built on it must display "Built with Muse Spark Open", and a company offering models as a service with more than $20 million of revenue in any twelve months needs a separate written agreement with Meta. Internal use and features inside a company's own product are outside that clause. VentureBeat, reporting the release on 16 October, notes that Meta's own table puts the model ahead on some tool-use tests and behind leading Chinese open models on others; those are Meta's figures and no score is stored here.
Key facts
Maker: Meta Platforms (Meta Superintelligence Labs) Released: 15 October 2026 Size: 816B total parameters, 42B active Context: up to 1M tokens Inputs: text, image, video, audio; output text Licence: Muse Model License (separate agreement for model-as-a-service providers above $20M revenue) Weights: BF16 and NVFP4 on Hugging Face, gated access
Benchmarks & AI stats
| Benchmark | Official | Community avg |
|---|---|---|
| Artificial Analysis Intelligence Index v5.0.1llm | 25 index | - |
“Official” values are editor-approved and feed the ranking. “Community avg” is the mean of member submissions (shown for transparency; it never affects the ranking until an editor approves a value).
Sign in to add a benchmark score for this model.
Categories & features
- United States
- Artificial Intelligence - AI
- Machine Learning
- Natural Language Processing
- Speech Recognition
Trust Score
An earned signal from verification, reviews, awards, transparency and engagement - the vendor can't buy it.
Updated 10/16/2026
Request a demo or quote from Muse Spark Open
Reviews
Write the first review of Muse Spark Open
Used it? Your experience helps other buyers decide.
Write a reviewQuestions & answers
No questions yet. Be the first to ask about Muse Spark Open.
Muse Spark Open alternatives
Similar listings buyers compare against this one.
Open-weight agentic coding model from IQuestLab with a 524,288-token context, for teams that self-host
NVIDIA's open-weights vision-language model, released 15 October 2026: 120B parameters with 12B active, a context of up to 1M tokens, and use in reasoning, coding, tools, documents and video.
Diffusion language model from Inception for enterprise teams building real-time voice agents, announced in September 2026
Mistral's first time-series forecasting models, released 15 October 2026 in four sizes from 5M to 209M parameters: open weights, probabilistic multivariate forecasts and CPU inference.
Mid-tier Anthropic model for everyday work, coding and document creation, with a 1 million token context window
Google's frontier model, announced in September 2026 and released first to selected cyber defenders
OpenAI's mid-tier GPT-6 model with a 1.05-million token context, image input and multi-agent support in beta
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- More United StatesBack to the ranked category
- Write a reviewHelp other buyers decide
- CompaniesAgencies, consultancies and IT service providers, ranked by verified reviews.
- ProductsSoftware and SaaS with pricing, features, integrations and alternatives.
- ArticlesGuides, comparisons and research from the editorial desk and the community.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.