Pricing check · 26 Oct 2026

AI Coding Assistants Pricing & Token ROI Calculator

Four coding seats, one workload. The total is the published subscription. A prompt is not a token, and this page does not invent one.

Price the seats Pricing Watch

Prices come from official vendor pages checked 26 Oct 2026. This is a planning calculator, not an invoice. Affiliate disclosure.

Embed this calculator · All embeds and badges

Verified pricing

ToolCheckedList priceWhat the meter is
Cursor26 Oct 2026Free $0 · Plus $8 · Premium $30 · Super $100 · Ultra $200 · Teams from $30/userWeekly shared pool: Premium 4× Plus, Super 15× Plus, Ultra 50× Plus. Third-party models are extra at API list price. No annual discount on the page.
Claude Code26 Oct 2026Pro $20/mo or $200/year · Max 5x $100/mo · Max 20x $200/moClaude Code is included on paid plans and shares the chat pool. No fixed message count. Max is monthly only.
GitHub Copilot26 Oct 2026Free $0 · Pro $10 · Pro+ $39 · Max $100 · Business $19/user · Enterprise $39/userAI credits: Pro 1,500, Pro+ 7,000, Max 20,000, Business 1,900, Enterprise 3,900. Extra credits are $0.01. Free includes 2,000 completions a month.
Windsurf26 Oct 2026Free $0 · Pro $20 · Max $200 · Teams $40/full seat ($80 minimum) · Enterprise customDaily and weekly quota. Prompt credits were retired in March 2026. Extra usage is API list price. Existing Pro subscribers can keep $15.

Sources: Cursor pricing, Anthropic pricing, Claude plan guide, GitHub Copilot plans, Windsurf pricing, Devin self-serve plans, Devin quota migration. Windsurf's pricing page is titled as Devin plans.

Team and solo cost calculator

Set the seats and the heavy prompts each developer runs in a day. The band is a planning label, not a measured burn rate: 5–25 is light, 26–60 is steady, and 61–100 is heavy. Frontier steps the published pool up where the vendor says harder models draw more of it. Overage stays at $0 unless the vendor publishes a price per prompt. None of these four do.

Team size
Daily agentic prompts per developer
Model tier
Billing cadence

The subscription totals load in this browser.

How the call is made

  • Keep is the lowest published subscription whose plan actually includes the model class you selected.
  • Test is another seat at or under 2.5× that price. It is worth a trial if the editor or the terminal is the job, not because the quota is proven to last.
  • Cut means do not also pay for that seat at this workload. It is overlap, or the published price is the wrong size.
  • Cursor's included pool is the X Subscription allowance. Frontier models from other labs are not in that price, so Cursor is not the Keep seat when Frontier Heavy is selected.
  • Annual billing changes Claude Pro only, from $20/month to $200/year. Max, Copilot, Cursor, and Windsurf have no lower annual list price on the pages checked.

Questions

Is Cursor Pro $20 enough or do I need Ultra $200?

Pro at $20 is not on cursor.com/pricing as of 26 Oct 2026. The page lists Free at $0, Plus at $8/month, Premium at $30/month, Super at $100/month, and Ultra at $200/month. Ultra is the power-user pool, published as 50× Plus. Premium, at $30, is the plan Cursor describes for new AI users, with a pool of 4× Plus. Super, at $100, is described for daily agent users and is 15× Plus. Third-party frontier models are not included in those prices. They bill at API list price. A legacy Cursor plan can continue: xAI's subscription docs, checked for this site on 13 Oct 2026, say existing Cursor plans continue and switching is optional. This calculator prices the plans a new buyer can select.

How much does Claude Code cost per month on Max vs API?

Claude Code is included with paid Claude plans and draws from the same usage pool as chat. Pro is $20/month, or $200/year ($17/month equivalent on Anthropic's pricing card). Max 5x is $100/month and Max 20x is $200/month. Max is monthly only. Anthropic's help center describes Max 5x as 5× Pro usage per 5-hour session and Max 20x as 20×. There is no fixed message count. After a limit, paid plans can continue on usage credits at standard API rates. The pricing page lists Sonnet 5.5 at $2 per million input tokens and $10 per million output tokens, and Opus 5.5 at $4 and $20. Max also includes monthly API credits ($100 on 5x, $200 on 20x) that Anthropic says do not apply to Claude Code. This calculator does not turn a prompt into tokens, so the API path is not added as a dollar overage.

What counts as a Premium Request in GitHub Copilot Pro vs Pro+?

GitHub's plans page checked 26 Oct 2026 meters GitHub AI Credits, not a unit called a premium request. Copilot Pro is $10/month with 1,500 total monthly AI credits (1,000 base plus 500 flex). Copilot Pro+ is $39/month with 7,000 (3,900 base plus 3,100 flex). Copilot Max is $100/month with 20,000. Copilot Free includes 2,000 completions a month and an AI-credit allowance the plans page does not number. Business is $19/user/month with 1,900 credits, and Enterprise is $39/user/month with 3,900. Usage past the pool is $0.01 per AI credit. Code completions stay unlimited on paid plans. The page does not say how many credits one agent prompt uses, so this calculator does not convert prompts into credits.

Windsurf Pro $15 vs Cursor Pro $20: how do prompt credits compare?

They are not the same unit, and neither sticker is the current public list price for a new buyer. windsurf.com/pricing, checked 26 Oct 2026, lists Pro at $20/month and Max at $200/month. Devin's quota docs say an existing Pro subscriber at $15/month keeps that price, and that March 2026 replaced prompt credits with a daily and weekly quota. Add-on prompt credits were converted at $0.04 per credit. Cursor's $20 Pro tier is not on cursor.com/pricing. The current individual rungs are Plus $8, Premium $30, Super $100, and Ultra $200. Windsurf Pro includes frontier models inside the quota. Cursor's included pool does not include third-party frontier models. Neither vendor publishes a prompt-credit balance you can line up one for one.

Which AI coding stack is cheapest for a 5-person engineering team?

At 20 prompts a day per developer, balanced models, and monthly billing, the published subscriptions are GitHub Copilot Pro at $50, Claude Pro at $100, Cursor Premium at $150, and Windsurf Teams at $200. Copilot Pro is the lowest published subscription, so the calculator marks it Keep, Claude Test, and Cursor and Windsurf Cut for that profile. That is the subscription, not a measured token bill. Copilot Business at $19/user ($95 for five) has a larger credit pool than Pro, and it is the org plan if you need centralized policy. Move the sliders for a different workload. Five heavy frontier seats are a different total.

Read the lane split in Claude Code vs Cursor vs Copilot, then run the Stack Audit if these seats are already on the card. Credit clocks are on Price changes.

Switch links: Cursor, Claude, GitHub Copilot, Windsurf.

Independent math: our Keep/Test/Cut formulas are open-source in the page JS. Some outbound switch links may earn a partner commission at no cost to you. Affiliate disclosure.

Running a 5+ seat team? Get a Custom AI Stack & Seat Audit · See it on the pricing page