Two of these three publish a number. One doesn't.
Claude Enterprise is $20 a seat per month, billed annually — but that seat fee only covers access. Actual usage bills separately at API rates, so the real invoice looks more like a cloud bill than a SaaS subscription, and it can land well above $20/seat once a team is actually using it. Gemini Enterprise starts at $21/seat. ChatGPT Enterprise still routes every buyer to a sales call and won't put a number on a page — which, in practice, usually means the price depends on how much leverage you walk in with.
What you get for the seat fee also isn't identical. Claude's version leans hard into the admin and compliance side: SCIM provisioning, audit logs, configurable data-retention windows, role-based access — the checklist a security team asks for before they'll approve any AI vendor. That's arguably worth more than the $1/seat gap to Gemini suggests.
| Published seat price | Pricing model | |
|---|---|---|
| Claude Enterprise | $20/seat/mo | seat + metered API usage on top |
| Gemini Enterprise | $21/seat/mo | seat-based, ties into Workspace |
| ChatGPT Enterprise | not published | sales-negotiated, credits/usage pools |
The thing that actually decides your total spend isn't the seat price — it's usage metering. A team of 50 people lightly using Claude might land near $1,000/month. The same team running agents and long-context research sessions could double or triple that on API charges alone, invisible until the first real invoice.
If procurement wants a number to plan around before any calls happen, Claude and Gemini both give you one. If your team already lives in Google Workspace, Gemini Enterprise's integration probably beats the dollar saved elsewhere. If you need the deepest admin/compliance tooling and can stomach usage-based billing, budget for Claude and then double your seat-price estimate before you present it internally.