Meta has officially launched its next-generation artificial intelligence model, Muse Spark, marking the first achievement of its Superintelligence Labs division. Previously, CEO Mark Zuckerberg invested approximately $14 billion to advance Meta's AI transformation, bringing in Alexandr Wang to lead relevant teams while accelerating R&D through large-scale hiring and aggressive talent acquisition. Developed on Meta's rebuilt AI tech stack encompassing entirely new infrastructure, model architectures, and data pipelines, Muse Spark introduces a "deep thinking" mode capable of orchestrating multiple agents to collaboratively execute tasks. It enhances response quality by integrating medical consultation data from roughly 1,000 doctors and adds e-commerce recommendation features that convert platform content into shopping suggestions. In capability comparisons, Meta positions its deep-thinking mode alongside Gemini 3.1 and GPT-5.4 Pro. The company states that the model has undergone extensive safety testing and demonstrates strong refusal capabilities regarding high-risk queries. Muse Spark is now live on the Meta AI platform and associated applications. Zuckerberg describes it as the first milestone toward achieving personal superintelligence, noting that additional models will be released in the future, some of which are planned for open-sourcing.