Meta Unveils ‘Muse Spark’ AI Model to Strengthen Its Superintelligence Ambitions

In a major push to regain momentum in the artificial intelligence race, Meta Platforms has introduced its first AI model developed by its high-profile superintelligence team—marking a significant step in its long-term AI strategy.

Meta Launches Muse Spark AI Model

Meta has officially unveiled “Muse Spark,” a new artificial intelligence model that represents the first output from its ambitious superintelligence division. This team was formed last year with substantial investments to compete with leading AI players globally.

The launch comes at a time when major tech companies are under intense pressure to demonstrate returns on their massive AI spending. With Muse Spark, Meta aims to showcase its progress and re-establish itself among the top innovators in the AI space.

Following the announcement, Meta’s shares surged by nearly 7%, reflecting investor optimism about the company’s AI direction.

Massive Investments to Catch Up in AI Race

Meta’s renewed focus on AI has involved significant financial commitments. The company made headlines with a reported $14.3 billion deal to bring in Alex Wang, the CEO of Scale AI, to lead its superintelligence efforts.

In addition, Meta has reportedly offered highly competitive compensation packages—some reaching hundreds of millions of dollars—to attract top engineering talent. These efforts highlight the company’s determination to close the gap with industry leaders such as Google, OpenAI, and Anthropic.

The initiative follows a relatively underwhelming reception to Meta’s earlier Llama 4 models, prompting the company to rethink its strategy and accelerate innovation.

What is Muse Spark and ‘Avocado’ Series

Muse Spark is the first model in a new internal series of AI systems codenamed “Avocado.” These models are designed to push the boundaries of what is often referred to as superintelligence—AI systems capable of surpassing human cognitive abilities in certain tasks.

While Meta has not disclosed the exact size or computational scale of Muse Spark, it described the model as “small and fast by design.” Despite its relatively compact structure, the AI is capable of handling complex queries in areas such as science, mathematics, and healthcare.

This approach suggests Meta is prioritizing efficiency and usability over sheer model size, at least in the early stages of development.

Limited Release and Strategic Rollout

Unlike its previous open-source approach with Llama models, Meta has taken a more controlled release strategy with Muse Spark. The model is currently available only through a “private preview” to select partners.

Initially, it will be accessible via the Meta AI app and website, which have so far seen limited adoption. However, the company plans to expand its integration across its ecosystem in the coming weeks.

Muse Spark is expected to replace existing AI systems powering chatbots on platforms such as WhatsApp, Instagram, and Facebook, as well as Meta’s smart glasses.

Performance Compared to Rivals

Independent evaluations indicate that Muse Spark is competitive with leading AI models in certain areas, particularly language processing and visual understanding. However, it still lags behind competitors in domains like coding and advanced reasoning.

The model reportedly ranked fourth on a comprehensive AI performance index compiled by Artificial Analysis, suggesting that while it is promising, there is still room for improvement.

Meta CEO Mark Zuckerberg had previously tempered expectations, stating that the initial models would demonstrate progress rather than immediate dominance.

Future Improvements and Roadmap

Alex Wang acknowledged that Muse Spark is still evolving and has “rough edges” that will be refined over time. Larger and more advanced versions of the model are already under development.

Meta also indicated that it plans to release some future models more openly, potentially returning to its earlier strategy of sharing AI technology with the broader developer community.

The company aims to continuously improve its models throughout the year, pushing the boundaries of AI capabilities and staying competitive in a rapidly evolving field.

Monetisation Through AI-Powered Features

With Muse Spark, Meta is also offering a glimpse into its monetisation strategy. The company is exploring ways to integrate shopping features directly into its AI chatbot.

These features would allow users to discover and purchase products seamlessly within conversations, creating new revenue streams. By embedding commerce into AI interactions, Meta hopes to capitalise on its massive user base.

Leveraging Meta’s Massive User Base

Meta’s biggest advantage lies in its vast ecosystem, which includes over 3.5 billion users across its platforms. By integrating AI into everyday tasks, the company aims to boost engagement and retain users.

Muse Spark is designed to assist with a wide range of activities, such as:

• Estimating calories from food images
• Visualising objects in real-world settings
• Planning trips and daily tasks

These features demonstrate how AI can be seamlessly incorporated into daily life, enhancing user experience and engagement.

Introduction of Contemplating Mode

One of the innovative features introduced with Muse Spark is “Contemplating Mode.” This functionality allows the AI to run multiple agents simultaneously, enhancing its reasoning capabilities.

This feature is comparable to advanced modes offered by competitors, such as Gemini Deep Think and GPT Pro. It enables the AI to handle complex, multi-step tasks more effectively.

For example, users can plan a family vacation by assigning one AI agent to create an itinerary while another searches for suitable activities, making the process more efficient and comprehensive.

A Step Forward in AI Competition

The launch of Muse Spark marks an important milestone for Meta as it seeks to reassert its position in the AI landscape. While the model is still evolving, it demonstrates the company’s commitment to innovation and long-term growth.

As competition intensifies among tech giants, the success of initiatives like Muse Spark will play a crucial role in shaping the future of AI and determining which companies lead the next wave of technological advancement.