Meta Releases Open Multilingual MoE Model Muse Spark Open
Understand the story
On October 15, 2026, Meta released Muse Spark Open, an open-weight multimodal mixture-of-experts (MoE) model, via the Hugging Face Blog. The model has roughly 816B total parameters with about 42B activated per token, supports inputs in four modalities — text, image, audio, and video — offers a context window of up to 1M tokens, and was trained for agentic coding scenarios; users can obtain the model weights on the Hugging Face platform. In the early hours of October 16, the AWS Machine Learning Blog announced that the model had landed on Amazon SageMaker JumpStart, released as open weights under the Muse Model License, noting that inputs support text, images, documents, audio, and audio-video, with text output. The Verge subsequently reported that Meta called it the strongest open-weight American model to date, excelling at long-horizon tasks, complex software engineering, and multi-step reasoning; the report also said Meta partnered with Apple to optimize the model so it can run on desktop via the latest MLX framework on two Apple M5 Ultra Mac Studios, with Apple also publishing a technical article about MLX and executives from both companies publicly exchanging pleasantries on X. On October 16, The Decoder reported that this marks Meta's return to the open-weight space after the Llama series, positioning it against Qwen, GLM, and Kimi; the same day, TechCrunch also reported that the model can run locally on a Mac, emphasizing that it was designed for complex coding and agent tasks. There are no reports yet of third-party benchmark results or community feedback.
AI overview of reporting · Updated 3 days ago
Developments
2 developments
- 10/16 01:43 · 1 reportsMeta Muse Spark Open Available on SageMaker JumpStartAWS Machine Learning Blog：Meta’s Muse Spark Open now available in Amazon SageMaker JumpStart
- 10/15 08:00 · 4 reportsMeta Releases Open-Source Multimodal MoE Model Muse Spark OpenHugging Face Blog：Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta
Reporting timeline
Explore different aspects of the story through its reporting.
- TechCrunch · AIMeta launches an open-weight AI model that runs on Macs
Meta released Muse Spark Open, an open-weight model it calls America's strongest open-weight model, with 816B parameters and a 1M-token context window, natively multimodal and designed for complex coding and agent tasks.
- The DecoderSelectedMuse Spark Open is Meta's answer to Qwen, GLM, and Kimi in the open-weight race
After the Llama series, Meta has returned to the open-weight space with Muse Spark Open, calling it the strongest American open-weight model.
- The Verge · AIMeta and Apple team up to make Muse AI models run better on Macs
Meta released Muse Spark Open, an open-weights model it calls the strongest American open-weights model yet, with 816B total parameters, natively multimodal, excelling at long-horizon tasks, complex software engineering and multi-step reasoning. Meta partnered with Apple to optimize the model, which can run on desktop via the latest MLX framework on two Apple M5 Ultra Mac Studios; Apple also published a technical article on MLX, and executives from both companies exchanged public pleasantries on X.
- AWS Machine Learning BlogMeta’s Muse Spark Open now available in Amazon SageMaker JumpStart
Meta's coding-optimized open-weight model Muse Spark Open is now available on Amazon SageMaker JumpStart. The MoE model has 816B total parameters with 42B active per token, supports a 1M-token context window, takes text, image, document, audio, and audio-video input with text output, and is open-weight under the Muse Model License.
- Hugging Face BlogSelectedWelcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta
Meta has released Muse Spark Open, an open-weight multimodal mixture-of-experts model with roughly 816B total parameters and about 42B activated per token. It supports text, image, audio and video inputs, offers a context window of up to 1M tokens, and was trained for agentic coding.
Story heat trend
Current heat3·Comparable peak43（10/16 05:00）·Comparable change over 24 hours -89%
The trend compares the same continuously observed sources, which may cover fewer sources than the current heat score. Hover or select the chart to inspect each hour; use the left and right arrow keys to navigate.