Muse Spark 1.2

Muse Spark 1.2 is Meta's multimodal reasoning model designed for complex agentic tasks, accepting text, images, video, audio, and PDF documents with a 1-million-token context window.

Strengths

Handles multi-step reasoning and tool use across multiple input modalities, enabling agents to process diverse document types and visual information in a single request.

Best for

Building agents that need to reason over long documents, video/image analysis, and tool orchestration — an upgrade from Muse Spark 1.1 with expanded capabilities.

Limitations

Focused on reasoning and agentic workflows rather than general-purpose chat; not the right choice if you need a lightweight or open-weight model.