What ChatGPT, Copilot, Claude and Gemini cost a business
Cost
Short answer: A business seat costs roughly the same everywhere, around 20 to 30 dollars per user per month. What actually sets them apart is what you already pay for, how many seats you must buy, and what it costs when you build something yourself instead of buying seats.
A seat per employee
List price per user per month, excluding VAT.
| Service | Annual | Monthly |
|---|---|---|
| Microsoft 365 Copilot Business | $21 | not confirmed |
| Microsoft 365 Copilot, Enterprise | $30 | $31.50 |
| Claude Team | $20 | $25 |
| ChatGPT Business | $20 or $25 | $25 |
| Gemini in Google Workspace | Included in the subscription | |
OpenAI blocks automated access to its pricing pages, so the ChatGPT figures come from 2 independent price trackers. They disagree on the annual Business price, so both numbers are shown. Microsoft’s US page lists Copilot Business at $21 with annual billing; we could not confirm its monthly price.
What you already pay for
Gemini comes with Google Workspace. If you have Workspace, you are already paying for it. We do not quote a figure, because Google’s pricing page showed different campaign prices on two visits the same day.
Copilot is an add-on. It needs a qualifying Microsoft 365 licence underneath, and Copilot Business covers up to 300 users. The simpler Copilot Chat comes at no extra cost.
Claude Team is sold to teams of 2 to 150 people. ChatGPT Business sits outside both office suites, so it comes on top of what you pay Microsoft or Google.
When someone needs more
Everyone has pricier tiers for heavy users. Claude Team has premium seats at $100 a month with annual billing, and Claude Max starts at $100. ChatGPT Pro costs $200 according to both trackers. Giving this to everyone rarely pays off. Giving it to 2 or 3 people often does.
When you build your own
When you wire a model into your own systems, you pay per , not per user. That is where the differences get large. Dollars per million tokens, from the vendors’ own pages:
| Model | Input | Output |
|---|---|---|
| GPT-6 Astra | $10 | $50 |
| Claude Fable 5.1 | $10 | $50 |
| Claude Opus 5 | $5 | $25 |
| GPT-5.6 Sol | $4 | $20 |
| Claude Sonnet 5 | $2 | $10 |
| GPT-5.6 Terra | $2 | $12 |
| Gemini 3.1 Pro | $2 | $12 |
| Claude Haiku 4.5 | $1 | $5 |
| Gemini 3.8 Flash | $0.75 | $3.75 |
| GPT-5.6 Luna | $0.20 | $1.20 |
Two of these prices are temporary. GPT-5.6 Sol is a promotional price through at least 21 November 2026, and Gemini 3.8 Flash doubles from 1 January 2027. Gemini 3.1 Pro is still in preview.
Nor is a token the same amount of text everywhere. Anthropic says Claude 4.7 and later produce about 30 percent more tokens for the same text, so the price per token does not tell the whole story.
Requiring data to be processed in a specific region adds a surcharge. OpenAI charges 10 percent extra for regional processing on newer models, and Anthropic charges 1.1 times the price for US-only processing.
What this comes to for your volume, you can work out in the cost calculator.
Seats or API?
Seats fit when employees use an assistant day to day. The API fits when AI goes into a process, such as case handling or customer service, where volume drives the price. Many end up with both. Which assistant fits you is covered in the guide to ChatGPT, Copilot and Gemini.
Sources
All prices retrieved on 14 September 2026.