Meta has introduced Muse Spark, the first model in its new Muse family developed by Meta Superintelligence Labs, marking a significant step in its push toward advanced artificial intelligence systems.

Muse Spark is a natively multimodal reasoning model designed to handle complex tasks across text, images and tools. It incorporates capabilities such as visual chain-of-thought reasoning, tool use and multi-agent orchestration, positioning it as a next-generation AI system aimed at “personal superintelligence.”

The model is available via Meta’s AI platform and app, with a private API preview being rolled out to select users.

A key highlight of the launch is “Contemplating mode,” a new feature that enables multiple AI agents to reason in parallel. This approach significantly enhances performance on complex tasks, allowing Muse Spark to compete with advanced reasoning systems such as Gemini Deep Think and GPT Pro.

According to Meta, Muse Spark demonstrates strong capabilities across multimodal perception, reasoning, health applications and agentic workflows. In benchmark testing, the model achieved 58 per cent on Humanity’s Last Exam and 38 per cent on FrontierScience Research tasks, highlighting its growing competitiveness in frontier AI development.

The launch also reflects Meta’s broader strategy of scaling AI across the full technology stack. The company is investing heavily in research, model training and infrastructure, including the development of its Hyperion data centre, to support future generations of larger and more capable models.

While acknowledging current limitations in areas such as long-horizon agentic systems and coding workflows, Meta said Muse Spark represents the foundation of a new generation of AI systems designed to deliver more autonomous, intelligent and personalised digital experiences.