Muse Spark 1.2 Contributor

Meta's multimodal reasoning model that processes text and images together, positioned as a cost-efficient tier relative to Muse Spark 1.3 Contributor with a 1-million-token context window.

Strengths

Handles multimodal inputs and reasoning tasks across text and images, supporting agentic workflows and tool use at a reduced tier.

Best for

Experimentation and prototyping with multimodal data when cost efficiency matters more than peak performance.

Limitations

An earlier iteration than Muse Spark 1.3 Contributor, so reasoning capabilities and model quality lag behind the newer release; not suited for production-critical agentic systems requiring top-tier performance.