"Muse Code is a terminal coding agent, like many of the other coding agents on the market," Wang said. "So you can install it with one command and then use it to take on complete software engineering tasks across a wide variety of use cases, planning changes, writing code, validating the results."
The coding agent is paired with Muse Spark 1.2, which Meta said was developed and trained alongside Muse Code. Wang said the joint development of the model and coding agent improves overall coding performance compared with the previous Muse Spark release.
Developers can access Muse Code through a pay-as-you-go pricing model that matches the API pricing introduced with Muse Spark 1.1, costing $1.25 per million input tokens and $4.25 per million output tokens. Meta is also offering a contributor tier that Wang said provides access at a significantly lower price for developers who choose to help improve the model.
"Wang added that Muse Code also has 'a contributor tier that gets you in at a significantly lower cost,' which he characterizes as being 'more than 10 times cheaper than than even the pay-as-you-go tier.'"
Meta said Muse Code is built on a coding-specific harness that enables developers to manage AI models during software projects. While developers can substitute different AI models or harnesses, Wang said the strongest performance comes from using Muse Code together with Muse Spark because the two technologies were designed alongside each other.
The company is also beginning to accept requests for zero-data retention, allowing enterprise customers to use Muse Code without their development data being retained for model improvement or other purposes. Wang described the capability as "a big enterprise feature that is important for folks."
Rather than positioning Muse Code around leading-edge performance, Meta said its strategy centers on offering competitive pricing while providing a coding agent integrated with its Muse Spark family of AI models.
This analysis is based on reporting from CNBC.
Image courtesy of Meta.
This article was generated with AI assistance and reviewed for accuracy and quality.
