Meta Launches Muse Code AI Coding Agent for macOS and Linux

Meta has launched Muse Code, a terminal-based AI coding agent for macOS and Linux. Available in beta and powered by Muse Spark 1.2, the latest coding-focused update to Meta's Muse Spark model family, the tool is designed to tackle complex software engineering tasks across large repositories.



After a one-command installation, Muse Code can plan changes, write code, and validate results while coordinating multiple persistent sub-agents that work in parallel in isolated environments. Its local event log records every model call, tool run, approval, and edit, allowing work to resume after a crash. Developers can also use built-in commands to generate a plan, stress-test it, or continue working toward a specified goal.


Muse Code uses a pay-as-you-go pricing model with two tiers. Standard pricing is $1.25 per million input tokens and $4.25 per million output tokens. Developers who opt into Meta's Contributor tier, allowing the company to use prompts and completions to train future AI models, pay just $0.10 per million input tokens and $0.20 per million output tokens—more than 90 percent less than the Standard tier. That option is limited to 60 requests per minute and 2.1 million tokens per minute, while the Standard tier allows 3,000 requests per minute and 4 million tokens per minute.

The launch marks Meta's first major push into agentic coding, joining a growing field of AI-powered developer tools that includes OpenAI's Codex and Anthropic's Claude Code.

Muse Spark 1.2 was co-trained alongside Muse Code, with significantly more training compute devoted to coding tasks and a broader range of training environments. It is also available through the Meta Model API with expanded global access.