Those Prices Undercut Anthropic's Sonnet 5 Model, Which Typically Costs $3 Per Million Input Tokens and $15 Per Million Output Tokens

Meta has announced an early beta of Muse Code, a terminal-based coding agent designed to compete with Anthropic's Claude Code and OpenAI's Codex. The tool is powered by Muse Spark 1.2, a new version of Meta's AI model that offers improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows. Like its rivals, Muse Code can handle software engineering tasks such as writing code, planning changes, and validating results, and it can manage multiple sub-agents to delegate tasks efficiently.

Meta's announcement includes sample projects like an interactive model of a photon sphere and a Plants vs. Zombies knockoff, plus a demo video of the tool building a webpage from an MP4 file. Affordability is a key differentiator: the tool uses pay-as-you-go pricing by default at $1.25 per million input tokens and $4.25 per million output tokens. In an interview with CNBC, Meta's Chief AI Officer Alexander Wang said the company would also offer a contributor tier at a significantly lower cost—$0.10 per million input tokens and $0.20 per million output tokens—requiring users to provide feedback, according to The Wall Street Journal.

Those prices undercut Anthropic's Sonnet 5 model, which typically costs $3 per million input tokens and $15 per million output tokens. If Muse Code performs well, the price difference could sway companies to choose Meta, a strategy that has already worked for Chinese AI firms like DeepSeek.