Meta Opens Developer Access to Muse Spark AI, Launches Upgraded 1.1 Version

Meta Platforms on Thursday unveiled long-anticipated developer access to its Muse Spark artificial intelligence model, simultaneously introducing an upgraded version called Muse Spark 1.1 — a move that puts the social media company in direct competition with Anthropic and OpenAI in the paid AI marketplace.

The company is billing Muse Spark 1.1 as its most powerful model yet for real-world programming and automated tasks, framing it as part of its broader goal of achieving what it calls “personal superintelligence.”

According to Meta, the upgraded model is capable of writing and fixing code, operating software and outside tools, processing text, images, and video, and completing complicated multi-step tasks with minimal human involvement.

The original Muse Spark model made its debut back in April, becoming the first text and reasoning AI model produced by the superintelligence team Meta assembled last year as it works to catch up with competitors in the fast-moving artificial intelligence race.

While the model launched in April, Meta had been quietly testing its Application Programming Interface — a digital connector that lets developers plug the model’s capabilities into their own software — with select partners in a private preview.

Now, developers across the United States can access Muse Spark through a public preview on Meta Model API, where they can experiment with prompts, compare results, and build prototype applications.

New users who sign up for the API will receive $20 in complimentary credits to try out the model before transitioning to a pay-as-you-go system. The pricing is set at $1.25 per million input tokens and $4.25 per million output tokens — higher than OpenAI’s entry-level GPT-5 mini and Anthropic’s budget-friendly Claude Haiku 4.5, but less expensive than Anthropic’s more advanced Claude Sonnet 4.6 offering.

Muse Spark 1.1 is now available in Thinking mode within the Meta AI app and on its website. The model is also expected to eventually take over from existing Llama models that currently power chatbots across WhatsApp, Instagram, Facebook, and Meta’s line of smart glasses.

Thursday’s launch follows an announcement earlier in the week when Meta said it was expanding its generative AI offerings by rolling out Muse Image — its first image-generation model from Meta Superintelligence Labs — across its suite of apps.