Meta has quietly rolled out Muse Spark, its most sophisticated AI model to date, featuring a unique 'contemplating' mode. Unlike earlier versions, Muse Spark deploys multiple AI agents to research and synthesize a consolidated answer, effectively mimicking a team of experts tackling a problem.
CNBC's Julia Boorstin reported that the model uses a multi-agent system to gather and analyze information before presenting a final response. This approach aims to improve accuracy and depth, especially for complex queries that require fact-checking or synthesis from multiple sources.
'Contemplating' mode allows the AI to pause, evaluate, and refine its output before finalizing—similar to how a human might deliberate. This step is designed to reduce errors and produce more thoughtful answers.
The release comes amid intensifying competition in the AI space, with other tech giants like Google and Anthropic also pushing advanced models. While Meta has not made a formal announcement, the update appears to be available to some users already.
Experts suggest that Muse Spark's agent-based architecture could set a new standard for reliability in generative AI, though questions remain about its computational cost and potential misuse.