Meta opens AI coding model to developers with Muse Spark 1.1 API launch

Meta has made its latest generation AI coding model broadly available to developers, marking another step in the company's effort to establish itself as a competitive force in artificial intelligence. The new Muse Spark 1.1 model is now accessible through a freshly introduced Meta Model API, which entered public preview this week for developers in the United States.

What's New in Muse Spark 1.1#

The latest iteration represents a significant advancement from the original model launched earlier this year. The upgraded version delivers enhanced capabilities tailored to development workflows: it can identify and resolve intricate programming bugs, provides improved support for multi-agent agentic systems, and introduces native multimodal abilities to process images, videos, and document-based content. These improvements address feedback Meta gathered from the developer community following the initial release.

Availability and Access#

The model is now live in Thinking mode within both the Meta AI app and Meta's web platform. The newly launched Meta Model API provides a pathway for engineers to integrate Muse Spark 1.1 into their own applications and development tools. New users will receive $20 in complimentary credits upon account creation. Previously, Muse Spark functionality was limited to Meta's direct platforms—including the Meta AI chatbot, Instagram, WhatsApp, and Meta's smart glasses.

Part of Meta's Broader AI Strategy#

The launch fits within Meta's larger effort to compete with established leaders like OpenAI, Google, and Anthropic in the AI landscape. The company has invested heavily in recent years—both through substantial capital expenditure and strategic hiring—to narrow the gap with competitors. This move comes alongside other recent releases, including Muse Image, a text-to-image generator that has faced scrutiny over its incorporation of Instagram user content.