GoKawiilMeta offers 95% price cut on Muse Spark model for users who share data
Meta introduced a new pricing tier for its Muse Spark model, built for coding and other agentic tasks, that slashes token costs by roughly 95% for developers who agree to let their prompts and outputs be used to train future versions. Standard pricing runs $1.25 per million input tokens and $4.25 per million output tokens, compared to just 10 cents and 20 cents respectively under the data-sharing plan. The move follows a scrapped internal program at Meta that tracked employee computer activity for training data, which was shelved in June after backlash.
GoKawiil's interpretation of the reporting above, not reported fact.
Training data drawn from real usage sessions has become a critical ingredient for improving agentic AI, since it captures the messy, unscripted workflows that synthetic data can't replicate. By financially incentivizing data contribution rather than mandating it, Meta is signaling how difficult it has become to collect this data ethically or voluntarily, especially as enterprises increasingly resist letting vendors train on their proprietary interactions. The strategy could set a precedent for other AI labs facing the same data bottleneck while under pressure to keep improving agent performance.
- Meta's contributor pricing cuts costs by about 95% for users who share prompts and outputs from Muse Spark.
- The offer follows the failed employee-monitoring data initiative Meta paused in June after internal pushback.
- Experts note large companies typically avoid data-sharing plans despite steep discounts, preferring privacy over savings.
Source: techcrunch.com — Tim Fernholz, 2026-09-03
Published there as: “Meta is paying to peek at how you use their latest AI model”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.