Meta has announced Muse Code, an advanced AI coding agent powered by the Muse Spark 2.1 model. Designed to operate directly within the terminal, the tool delivers complete workflow automation for developers while significantly undercutting industry market prices.

Built to process and comprehend complex codebases, Muse Code functions as an autonomous technical collaborator. Rather than simply generating snippets, the system manages full development lifecycles through key capabilities:

• Automated Code Generation: Writes new code based on project specifications.
• Project Analysis: Audits existing codebases to plan structural changes.
• Debugging & Verification: Identifies software bugs and executes final code verification.

To further accelerate software development, Muse Code features multi-agent coordination. This functionality allows the system to delegate subtasks to multiple specialized AI agents simultaneously, ensuring faster project completion.

Meta’s competitive strategy hinges on a drastically reduced token pricing structure compared to existing market solutions:

• Standard Rates: Set at approximately $1.15 per million input tokens and $3.92 per million output tokens.
• Developer Feedback Rates: Users who opt to share performance feedback receive heavily discounted pricing at $0.10 per million input tokens and $0.18 per million output tokens.

By lowering cost barriers, Meta aims to make AI-powered software engineering accessible for large-scale operations while competing directly against established AI development tools.

Filed in . Read more about AI (Artificial Intelligence) and Meta.