Estimate your likely costs based on expected usage
The cost calculator helps you roughly estimate what your workspace will consume per month. Useful ahead of a workshop, a rollout, or when you need to communicate a budget ballpark to leadership.
You enter your headcount and pick one of three fixed usage profiles (High usage, Balanced, Low usage). Each profile splits the workforce in fixed shares across three intensity tiers — power, medium, and light users — with preset monthly action counts (AI chats, text and document translations, image generations). For each action, the calculator applies a flat average credit value; these values are hardcoded into the calculator and independent of the per-model, per-token table on All models. There's no free-form field for actions per user, and the calculator doesn't add a platform base fee — the estimate comes purely from the projected credits.
For context: the per-seat base fee doesn't cover any AI usage — it comes on top of the usage estimated here. Your expected total cost is therefore the per-seat base fee (see Manage billing) plus the usage-based portion the calculator estimates. That applies equally to every action; premium models like Claude Opus or GPT-5 simply consume more credits per action than cheaper standard models (see Concept).
Important: this is an estimate, not a binding quote. Actual costs depend on:
Model choice — cheap models (GPT-5 Nano, Gemini 2.5 Flash) can be 10× cheaper than premium ones (Claude Opus, GPT-5.4 Pro)
Per-user intensity — heavy users typically consume 5–10× more than casual users
Feature mix — image generation and complex workflows push consumption up faster than pure text chat