Back to Daily Feed 
Meta Introduces Muse Spark 1.1 with API Access & Agentic Improvements
Must Read
Originally published on Simon Willison's Weblog by Simon Willison
View Original Article
Share this article:
Summary & Key Takeaways
- Meta has released Muse Spark 1.1, now with an official API.
- The model boasts significant improvements in agentic tool calling and computer use.
- Simon Willison highlights "Attractor States in Self-Conversation" as a notable feature.
- A new
llm-meta-aiplugin is available for CLI and Python access to the model. - The release includes an evaluation report detailing its capabilities.
Our Commentary
Meta's Muse Spark 1.1 with an API is a welcome addition to the open-source LLM landscape. Simon Willison's quick integration with llm-meta-ai is a testament to the power of community tooling. The "Attractor States in Self-Conversation" sounds like a wild concept; I'm genuinely intrigued by what happens when models talk to themselves. This is a good step for Meta in the agentic space.
View Original Article
Share this article: