Meta Launches Muse Spark 1.1 to Expand AI Coding Push

Meta has officially introduced Muse Spark 1.1, the latest version of its multimodal artificial intelligence model built to support agentic coding and enterprise software development. The release marks another step in Meta’s expanding AI portfolio as the company seeks to compete with coding-focused models already available from OpenAI and Anthropic.

The newest version follows the initial Muse Spark announcement in April and is designed to complete complex software engineering tasks with minimal human intervention. According to Meta, Spark 1.1 can perform multistep reasoning, coordinate digital workflows, manage enterprise processes, and assist developers in deploying features across business systems.

The launch comes as technology companies continue investing heavily in AI systems capable of automating software development, improving developer productivity, and reducing repetitive engineering work.

Meta Targets Enterprise Development With Competitive AI Pricing

Although Meta entered the AI coding market later than several rivals, the company is positioning Muse Spark 1.1 as a cost-effective alternative for organisations adopting AI-powered software development tools.

According to Reuters, Meta will price the model at $1.25 per million input tokens and $4.25 per million output tokens. Those rates place Spark 1.1 close to competing offerings, includingAnthropic’ss Claude Haiku 4.5 and OpenAI’s GPT-5.6 Luna.

Meta says Spark 1.1 is designed to process demanding agentic workloads while helping businesses automate software maintenance, resolve programming bugs, and simplify large-scale code migrations. These capabilities reflect the growing demand from enterprises seeking AI systems capable of completing complex engineering tasks with limited manual oversight.

"Muse Spark 1.1 delivers exceptional performance in personal agentic tasks that require planning and orchestration across a range of external apps and services," Meta said in a company blog post.

Over the past several years, Meta has steadily expanded its lineup of foundation AI models, investing heavily in technologies that support developers, creators, and enterprise customers.

Zuckerberg Signals More AI Models Are on the Way

The importance of the Muse Spark 1.1 launch was highlighted by Meta CEO Mark Zuckerberg, who returned to posting on X for the first time in three years. His previous post on the platform came in July 2023, shortly after Twitter was rebranded as X.

"a strong agentic and coding model at a very low price," Zuckerberg wrote, adding that the model was "strongest at agentic performance, tool use, and computer use."

Zuckerberg also hinted at additional announcements, saying there is “more to come soon,” suggesting Meta plans to expand its AI model portfolio.

The release continues an active week for artificial intelligence announcements across the technology industry. Earlier this week, Meta introduced its Muse Image AI image-generation model. Other notable launches included a new version of Grok from SpaceX AI andOpenAI’ss GPT-5.6 family of models, both announced on Thursday.

With major AI developers releasing new foundation models at a rapid pace, competition across the generative AI market continues to intensify. Companies are increasingly differentiating themselves through pricing, performance, enterprise capabilities, and specialised tools aimed at developers and business users.