AI cost calculator for teams
Three questions: how many developers, how heavily they use agents, and what you pay for today. The answer is your monthly cost now next to the same work on Standard Compute, based on how coding agents are really used.
Where the usage numbers come from
Each preset is the agent work of real developers: active coding‑agent users on our gateway from August to November 2026, about 60 people, priced at Claude Sonnet 5.5 list rates over 20 working days a month. Most of it is re‑read from the prompt cache, which is why the cost per token is low. The last column is what the same users actually paid on Standard Compute, before the plan or pay‑as‑you‑go fee: about 70% less, because routine steps run on cheaper models (some chosen by our router, some by the users).
| Preset | What it looks like | At Claude list prices | Paid on Standard Compute | Basis |
|---|---|---|---|---|
| Light | A few agent prompts a day, alongside normal coding | $80 | $15 | 25th percentile of active users |
| Typical | Agents are part of the daily routine: several tasks a day, some running unattended | $150 | $45 | median of active users |
| Heavy | Agents most of the day, often several in parallel, long sessions | $700 | $190 | 90th percentile of active users |
Anthropic’s own figure for enterprise deployments is about $13 per developer per active day and $150 to $250 a month, with 90% of users under $30 a day (Anthropic: Claude Code costs). Our median day is lower and our heaviest users are heavier, which is what you would expect from teams that came to us to get past seat limits. If your developers mainly run Opus, choose it under Adjust assumptions: it costs twice as much per token.
What your team gets on Standard Compute
The calculator compares cost. These are the controls that come with it.
No 5‑hour or weekly caps
Point Claude Code and Claude Desktop at Standard Compute and your team stops hitting 5‑hour and weekly caps: agents run until the monthly budget is used. Claude is served across several clouds, so one provider’s quota or outage doesn’t stop your work.
A budget for every person and app
Every person and app gets a fixed monthly budget that never overruns, and you can set spending caps per key and for the whole company. Optional top‑off from a prepaid company balance keeps critical agents running.
The cheapest model that can do each step
A single API gives you frontier and open‑source models, and it routes each step to the cheapest model that can do it and the cheapest provider fast enough to serve it.
Processing in Europe
Set a Europe rule for your company, a team or a single key, and every request, including retries and fallbacks, is served only by approved European inference endpoints. If no European endpoint can serve it, the request is refused before your prompt is sent. The rule is free on every company plan.
Sign‑in with work accounts
Connect Microsoft Entra ID or any SAML identity provider so your people sign in to Standard Compute with their work accounts. Employees sign in to Claude Desktop and Claude Code with their Microsoft work account, through our EU gateway, so there are no API keys to hand out.
Prices used
Taken from each vendor’s own pricing and help pages, all checked on 7 November 2026. All prices in US dollars, before tax.
| Item | Price | Source |
|---|---|---|
| Claude Team, Standard seat | $20 per person per month billed annually, $25 monthly. 2 to 150 people | claude.com/pricing |
| Claude Team, Premium seat | $100 billed annually, $125 monthly | claude.com/pricing |
| Claude Team usage limits | Per member, a 5‑hour window and a weekly window, shared with Claude chat. A Premium seat has 5x the usage of a Standard seat | claude.com/pricing |
| Claude Enterprise, self-serve | $20 per seat per month billed annually, 20‑seat minimum, plus usage at API rates | claude.com/pricing |
| ChatGPT Business, Standard seat | $20 per user per month billed annually, $25 monthly. 2 to 200 people | openai.com/business/pricing |
| ChatGPT Business, Premium seat | $100 billed annually, $125 monthly | openai.com/business/pricing |
| ChatGPT Enterprise | Not published; contact sales | chatgpt.com/pricing |
| Claude Sonnet 5.5 | $2 input, $0.10 cache read, $2.50 cache write, $10 output per million tokens | Anthropic API pricing |
| Regional endpoints on Bedrock and Google Cloud | 10% over global endpoints; Microsoft Foundry US Data Zone 1.1x | Anthropic API pricing |
| Standard Compute plans per key | Starter $19, Economy $39, Standard $89, Pro $249, Pro Plus $499, Growth $999, Scale $2,499; budget equals price | standardcompute.com/pricing |
| Standard Compute company plans | Teams free, SSO $399 a month, Enterprise $1,499 a month. Pay as you go: usage + 7.5% on Teams, + 10% on Enterprise | standardcompute.com/pricing |
Cloud usage is priced at Anthropic’s list rates. Microsoft Foundry bills Claude at those rates; Amazon Bedrock and Google Cloud publish their own price pages, so check yours if you have a private offer. Claude Team can also turn on usage credits so members keep working past their limits; the calculator leaves them out.
Questions from engineering leads
- How much does Claude Code cost for a team of 20?
- It depends mostly on how heavily they use agents. For 20 developers using agents every day, Claude Team Premium seats cost $2,000 a month billed annually, and self‑serve Claude Enterprise about $3,400 (20 seats at $20 plus usage at API rates). On Standard Compute the same team comes to about $970 a month, with no 5‑hour or weekly limits.
- How much does one developer use with Claude Code in a month?
- Anthropic says that across enterprise deployments the average is around $13 per developer per active day and $150 to $250 per developer per month, with 90% of users below $30 per active day. On our gateway, the median active coding‑agent user does about $7 of agent work per active day at Claude Sonnet 5.5 list prices, roughly 22 million tokens, most of it read from the prompt cache. Over 20 working days that is about $150; a light user is about $80 and a heavy one about $700.
- Is a Claude Team Premium seat cheaper than paying for usage?
- Usually yes, as long as developers stay inside the seat’s limits. A Premium seat is a flat $100 a month billed annually, while a heavy developer’s agent work is worth about $700 a month at API list prices. For 20 heavy users that is $2,000 on Claude Team against about $4,100 on Standard Compute. The difference buys no 5‑hour or weekly limits: on a seat, a developer who reaches a limit waits for it to reset.
- What does Claude cost on Bedrock or Azure for 120 developers?
- You pay per token, so it scales with use. For 120 developers using agents every day that is about $18,000 a month at global endpoint list prices, and 10% more on regional endpoints such as EU‑only on Bedrock. On Standard Compute it is about $5,800, because active users on Standard Compute paid about 70% less than list prices for the same work, with routine steps on cheaper models (some chosen by our router, some by the users).
- What does Claude Enterprise cost per user?
- Anthropic lists self‑serve Claude Enterprise at $20 per seat per month billed annually, with a 20‑seat minimum, plus all usage at API rates. For anyone running agents the usage is the larger part: about $150 a month for a developer using agents every day.
- How much does ChatGPT Enterprise cost?
- OpenAI does not publish a ChatGPT Enterprise price; its pricing page asks you to contact sales. ChatGPT Business is listed at $20 per user per month billed annually and $100 for a Premium seat. You can enter your own contract prices under Adjust assumptions.
- How do we stop one engineer’s agents from using the whole AI budget?
- On Standard Compute every person and app gets a fixed monthly budget that never overruns, and you can set spending caps per key and for the whole company. Optional top‑off from a prepaid company balance keeps critical agents running.
More on limits in Claude rate limits for teams, and on controls in Standard Compute for companies.