Meta's new Muse Spark coding-agent model ties steep pricing discounts to explicit consent for using customer interactions to train future models.
Meta is offering an average 95% discount on its new Muse Spark model — built for coding and other agents — to users who agree to let the company use their usage data to train future models. The arrangement is explicit and opt-in rather than a default term buried in a ToS update.
The move formalizes a pay-with-data pricing structure at a moment when developer tools are increasingly monetized through the value of interaction data itself, not just API fees.
This turns data-sharing into a line-item pricing lever rather than a background policy, which could pressure competitors to either match the discount or defend premium pricing on privacy grounds. Enterprises building on Muse Spark need to weigh cost savings against exposing proprietary code and workflows to a third party's training pipeline.
The daily signal, curated. Get it in your inbox.
Subscribe on LinkedIn →