Meta Unveils Muse Spark 1.1 as It Pushes Into AI Coding Market

Key Takeaways

• Meta launched Muse Spark 1.1, its latest AI model optimized for coding and agentic AI tasks.

• The company introduced a paid developer API with usage-based pricing and $20 in free credits for new accounts.

• Meta is positioning the model as a competitor to offerings from OpenAI and Anthropic.

• The rollout marks a broader shift toward commercializing Meta's AI models through paid developer services.

What Happened

Meta Platforms on Thursday introduced Muse Spark 1.1, the latest version of its artificial intelligence model designed for software development and autonomous AI applications.

The updated model is being released through a public developer preview, allowing developers to register for API access via Meta's developer portal. Existing partners already have access, while additional users will be admitted through a waitlist as capacity expands.

Unlike earlier limited releases, Muse Spark 1.1 represents a broader commercial launch that includes usage-based pricing for developers.

Alexandr Wang, Meta's Chief AI Officer, said the company designed the model to improve coding performance and support AI agents capable of completing complex, multi-step tasks.

Why It Matters

The launch reflects Meta's effort to strengthen its position in the rapidly expanding market for enterprise AI models, where OpenAI, Anthropic, and Google have established strong developer ecosystems.

By introducing paid API access, Meta is also expanding its AI monetization strategy beyond advertising. The move provides developers with direct access to proprietary models while creating a potential recurring revenue stream tied to AI infrastructure usage.

The announcement comes as investors continue to scrutinize Meta's multibillion-dollar investments in AI infrastructure and advanced model development.

Key Details

Meta said every new API account will receive $20 in free usage credits.

After those credits are exhausted, developers will pay:

• $1.25 per one million input tokens

• $4.25 per one million output tokens

According to Wang, the pricing is intended to remain competitive with other leading AI model providers while encouraging broader adoption.

Muse Spark 1.1 has been optimized for coding workflows and integration with widely used developer tools and AI agent frameworks. Meta said the model is currently available through its own infrastructure rather than third-party AI model marketplaces.

The company also confirmed it is training a more advanced model, internally code-named Watermelon, while continuing development of an open-source variant of Muse Spark. No release timeline was announced.

Earlier this week, Meta also introduced Muse Image, an AI image-generation model aimed at creators and advertisers, expanding the company's growing family of proprietary AI products.

Market Impact

The launch highlights Meta's increasingly commercial approach to artificial intelligence as competition intensifies among major AI developers.

While Meta previously emphasized open-source releases through its Llama family of models, the company is now combining open-source development with proprietary paid services designed for enterprise customers and software developers.

The strategy aligns Meta more closely with the business models adopted by OpenAI and Anthropic, both of which generate revenue through API access and enterprise AI services.

Conclusion

Muse Spark 1.1 marks another step in Meta's effort to establish itself as a leading provider of enterprise AI models. By combining advanced coding capabilities with a paid developer platform, the company is broadening its AI business beyond research and consumer applications.

As demand for AI coding assistants and autonomous software agents continues to grow, the success of Meta's commercial AI strategy will likely depend on developer adoption, model performance, and its ability to compete against established rivals in an increasingly crowded market.