The Pulse
ai& makes Meta’s Muse Spark Open available in Japan
ai& has made Meta’s Muse Spark Open available through its ai& Inference platform in Japan. The model runs on ai&’s infrastructure in Japan, with usage-based pricing and input and output excluded from model training.
AI.info Team · Written with AI · fact-checked ·
Availability in Japan
ai& has made Meta’s Muse Spark Open available through its ai& Inference platform in Japan. In a Japanese-language announcement on Nov. 10, ai& said users can try the model through an OpenAI-compatible API or a playground.
The model ID is meta/muse-spark-open, and the model can be called from existing applications through the compatible API. The service uses usage-based pricing.
Model specifications
ai& describes Muse Spark Open as Meta’s first open-weight Muse model and says it is designed for coding and long-running agents. The release identifies it as a native multimodal mixture-of-experts model, with 816 billion total parameters and 42 billion active parameters.
Meta positions the model as the highest-performing open-weight model from the United States to date, according to ai&’s announcement. The release says Muse Spark Open supports a 1 million-token context window and adjustable inference strength, and is suited to tool use and agent workflows.
Processing and data
All processing takes place on ai&’s infrastructure in Japan. Inputs and outputs are not used to train the model.
The announcement presents domestic processing as part of the service’s setup, while making the model accessible through an API compatible with OpenAI’s format.
Service plans
ai& says it aims to make newly released models available in Japan as quickly as the same day, and plans to add further public models to ai& Inference. The company describes itself as a Japan-origin AI business centered on inference, with work spanning semiconductor and computing infrastructure optimization, model execution and additional training, and application development and operations.
ai& runs its own data centers in Japan and develops and operates infrastructure and technology for using AI in services and work. The company also offers services in areas including language, video generation and scientific research through ai& Inference.