Meta Unveils Initial AI Model from Breakthrough Research Group

Meta Platforms Inc. Unveils Muse Spark, Its First Closed-Source AI Model

Meta Platforms Inc. has debuted its latest artificial intelligence model, Muse Spark, marking the company's first closed-source AI model since Chief Executive Officer Mark Zuckerberg embarked on a multibillion-dollar overhaul of the company's AI organization to keep pace with rivals.

The highly anticipated model, created by Meta Superintelligence Labs, the new team of expensive AI researchers helmed by Chief AI Officer Alexandr Wang, will power the Meta AI chatbot. This marks a pivot from the company's prior open-source strategy, with the design and code of Muse Spark remaining private. The model is the first major test for Zuckerberg's new AI lab, known as MSL.

As part of a $14 billion investment into Scale AI last year, Zuckerberg recruited Wang to lead the new AI lab. This investment was made after a series of setbacks left Zuckerberg frustrated that the company wasn't keeping pace with competitors like OpenAI, Anthropic PBC, and Alphabet Inc.'s Google. Since then, Zuckerberg has spent billions on AI talent and committed tens of billions more on infrastructure like data centers to catch up.

Read also: Google Pixel Watch Leak: Diver's Discovery Sparks Speculation

Early Results and Comparison

While Muse Spark may not be as capable in some areas as its competitors, company executives see it as an early milestone on their journey. The model was built over nine months and is considered a refresh of Meta's AI strategy, which was previously focused on its open-source Llama models. Wang is a proponent of closed models, and while Meta still plans to build open-source ones in the future, it is also considering selling API access to Muse Spark.

The Meta AI chatbot will remain free for users, but the company is considering implementing subscription fees in the future. Muse Spark was trained using several third-party open-source models, including Qwen from the Chinese tech giant Alibaba Group Holding Ltd., as well as ones from OpenAI and Google. This move has raised concerns about China's position in the fast-moving AI race, citing national security risks.

Read also: Chinese Startup Claims AI Coding Capabilities Exceed Those of GPT-5.5: A Detailed Examination

Muse Spark offers several levels of reasoning: "Instant" mode, "Thinking" mode, and "Contemplating" mode, the last providing research-grade responses. The model is good at answering questions about science, health, and math but weaker at coding. It will be used to improve Meta's fleet of apps, from Instagram to Facebook to WhatsApp, and is already powering an early-stage shopping agent that aims to help users better search for products like clothing or furniture.