How can we help?

Token usage is calculated based on both input (prompts) and output (completions). Each model has a specific weight—for example, higher-compute models like Komen-Ultra consume more tokens per generation compared to Komen-Lite.

Yes, the Team and Pro Tier accounts allow for workspace sharing. You can invite collaborators via email and set specific permission levels (Admin, Editor, or Viewer) for each project.

By default, Enterprise and Pro Tier data is strictly isolated and never used for training foundation models. Free tier users can opt-in to help improve model accuracy, but all data is anonymized before processing.