Meta Unveils Muse Spark AI: Aims to Close Ground on OpenAI, Google
Meta launches Muse Spark, its first AI model from Superintelligence Labs.
Top Summary
- What happened: Meta has released Muse Spark, its inaugural AI model developed by Meta Superintelligence Labs (MSL).
- Why it matters: This marks Meta's serious push to compete with AI leaders like OpenAI, Anthropic, and Google after lagging behind.
- What changes for people: Muse Spark will power Meta AI assistant, enhancing everyday tasks and potentially impacting over 3.5 billion users.
- Who is affected: Users of Meta AI app, WhatsApp, Instagram, Facebook, and Meta's smart glasses will experience the AI's capabilities.
Meta's AI Revamp: Muse Spark Arrives
Meta has officially unveiled Muse Spark, its first AI model developed by Meta Superintelligence Labs (MSL). This launch signifies a major step in Meta's efforts to catch up with competitors in the rapidly evolving AI landscape.
Mark Zuckerberg established MSL under Chief AI Officer Alexandr Wang. Zuckerberg hired Wang in June 2025 after investing $14 billion in Scale AI. This followed frustrations that Meta wasn't keeping pace with OpenAI, Anthropic, and Google.
Inside Muse Spark: Features and Capabilities
Muse Spark is designed to be small and fast, focusing on reasoning through complex problems in science, math, and health. It currently powers the Meta AI assistant in the Meta AI app and on meta.ai.
The model boasts several reasoning modes: Instant, Thinking, and Contemplating. The latter offers research-grade responses. Shengjia Zhao, chief scientist at Meta MSL, highlighted the overhaul of Meta's entire AI stack to support scaling for “personal superintelligence.”
"Today we're releasing muse spark, the first model from MSL... Muse Spark is the result of that work, and now it powers Meta AI," said Wang.
Closed Model Strategy & Training
Unlike Meta's previous open-source approach with Llama, Muse Spark is a closed model. Its design and code will not be made public.
Muse Spark was trained using third-party open-source models, including Qwen from Alibaba, as well as models from OpenAI and Google.
Real-World Applications: From Shopping to Travel Planning
Muse Spark assists with everyday tasks. This includes calculating meal calories from photos and superimposing images of objects in spaces.
The “Contemplating” mode uses multiple agents for enhanced reasoning. One agent can create travel itineraries, while another finds kid-friendly activities.
Meta AI with Muse Spark features a "Shopping mode" for buying clothes or decorating rooms. "Shopping mode draws from the styling inspiration and brand storytelling already happening across our apps, surfacing ideas from the creators and communities people already follow," Meta said.
Availability and Future Plans
Muse Spark is initially available on the Meta AI app and website. It will soon replace Llama models powering chatbots on WhatsApp, Instagram, Facebook, and Meta's smart glasses.
The Meta AI chatbot will remain free for now. However, the company is reportedly considering subscription fees in the future.
Playing Catch-Up in the AI Race
Meta acknowledges that Muse Spark isn't as capable as competitors like ChatGPT, Claude, or Gemini in some areas. Despite advancements in generative AI, Meta is yet to dominate the AI model market.
A Meta executive called the model "an early data point on our trajectory," indicating larger models are in development. The model was reportedly known internally as Avocado during development and was built over nine months.
What to Watch Next
Expect to see updates regarding larger AI models from Meta in the coming months. The performance of Muse Spark across Meta's platforms will be closely monitored to gauge its impact on user engagement and overall AI strategy.