Meta has introduced a new pricing structure for its Muse Spark model that rewards users for sharing their data. The model is designed to power coding tools and other AI agents.
Under the new system, companies can choose to let Meta use their prompts and outputs to help train future models. In exchange, they receive a steep discount on usage costs.
The numbers show just how large that discount is. Standard pricing runs $1.25 per million input tokens and $4.25 per million output tokens.
How the Pricing Works
Companies that agree to the contributor tier pay far less. Input tokens drop to 10 cents per million, and output tokens fall to 20 cents per million.
That works out to roughly a 95% reduction in cost. Meta describes the lower tier as a way to help developers test and scale projects where sharing data is acceptable.
The company has struggled to gather the kind of data needed to train agentic AI tools. Earlier this year, Meta launched a project to track how employees used their computers.
That effort drew criticism from staff over privacy concerns. Meta paused the program in June while it reviewed data security issues.
Meta did not respond to a request for comment on the new pricing model.
Why Companies Hold Back Data
User data plays a large role in improving AI agents. One developer told TechCrunch that gains in coding agent performance last year came largely from stored user sessions used in training.
But gathering this kind of data is harder outside of software engineering. Many professional workflows do not leave the same kind of digital trail that coding sessions do.
Arvind Narayanan, a computer science professor at Princeton, pointed out that large companies tend to avoid sharing their data with AI providers.
He noted that many businesses stick with enterprise plans billed by token use, even though subscription plans cost far less. The main difference between the two options is data retention and IT control.
Narayanan suggested Meta's new pricing could push companies to think more carefully about which data is worth protecting and which could be shared without risk.
The move also fits into a wider pattern of price competition among AI companies. Anthropic's newest Fable and Mythos models launched with lower costs for cached tokens.
OpenAI cut prices on its latest models at the end of July. Meta's contributor pricing adds another option to that shifting cost landscape.
For now, Meta has not shared how many companies have signed up for the discounted tier. The pricing guide remains available on its developer site.
