Executive Summary
Meta has released Muse Spark 1.1, a multimodal reasoning model, on the AI Gateway platform. Designed for complex agentic tasks, the model features a one-million-token context window and can process text, image, video, PDF, and audio inputs. Muse Spark 1.1 is built to plan and orchestrate work across various tools and services, simplifying development for sophisticated AI applications.
Key Takeaways
* Product Name: Muse Spark 1.1
* Primary Function: A multimodal reasoning model for building and executing agentic tasks.
* Context Window: 1 million tokens.
* Multimodal Inputs: Accepts text, image, video, PDF, and audio.
* Agentic Capabilities: Can plan and orchestrate workflows, operate as a main agent or subagent, and adapt to new tools without requiring examples.
* Advanced Features: Supports parallel tool calling, structured output, and built-in search with citations.
* Availability: Now available on AI Gateway by setting the model to `meta/muse-spark-1.1`.
Strategic Importance
This launch provides developers with access to a powerful, large-context multimodal agent via a unified platform, positioning Meta as a key provider of foundational models for complex, automated workflows.