Claude Code pricing runs through Anthropic’s paid Claude plans, starting with Pro at $20 billed monthly or $200 billed up front for a year. The other route is pay-as-you-go API tokens on a Claude Console account. Your chats and your coding sessions draw on one usage pool with a 5-hour window and a weekly limit, so a bigger plan buys capacity, not a better Claude Code.
The Free plan leaves Claude Code out. Max costs $100 and Ultra $200 per month, both billed monthly only. Anthropic’s Claude Code cost docs put average API spend at about $13 per developer per active day, so 2 active days on tokens (2 x $13 = $26) already cost more than a month of Pro.
Quick pricing answer
Pro at $20 per month billed monthly ($200 a year billed annually) fits a developer who codes a few days a week. Max suits all-day solo coding, and Ultra drops the 5-hour window for verified accounts. A dev team buys Team seats. Standard seats cost $20 per member per month, billed annually. Premium seats cost $125 per member per month, billed monthly.
Claude Code has no free tier, and the Team plan lists Trials as “No”. The caveat: the $100 to $500 in monthly API credits bundled with Max, Ultra and Team can’t pay for Claude Code, and usage past your limit bills at standard API rates. Checked November 2026, USD, US prices.
In This Article (15 sections)
Claude Code Pricing at a Glance
The cheapest way into Claude Code is Pro at $20 per month billed monthly, the largest individual plan is Ultra, Team seats run $20 to $125 per seat per month, and a Console account bills per million tokens.
| Plan | Monthly billing | Annual billing | Claude Code usage | Main limitation |
|---|---|---|---|---|
| Free | $0 | n/a | Not included | No Claude Code |
| Pro | $20 per month, billed monthly | $17 per month, billed annually ($200 up front) | 1x every 5 hours and every week | Fable only through usage credits |
| Max | $100 per month | Monthly only | 5x Pro every 5 hours and every week | The 5-hour window still applies |
| Ultra | $200 per month | Monthly only | 10x Pro every week, no 5-hour limit for verified accounts | Monthly billing only |
| Team Standard seat | $25 per seat per month, billed monthly | $20 per seat per month, billed annually | More than Pro, per seat | Teams of 2 to 150 only |
| Team Premium seat | $125 per seat per month, billed monthly | $100 per seat per month, billed annually | 5x a Standard seat | Fable capped at 50% of weekly limits |
| Enterprise | Annual only | $20 per seat per month plus usage at API rates | No per-seat usage limits on usage-based plans | 20-seat minimum |
| Claude Console (API) | Usage-dependent | n/a | Billed per token; workspace spend limits and organization rate limits apply | Every token is metered |
Sources: Anthropic’s Claude pricing page and the Claude API pricing docs, checked November 2026.
These are US prices in US dollars, read on November 7, 2026; they exclude applicable tax, and Anthropic says prices and plans are subject to change at its discretion.
Capacity is stated as multiples of Pro (1x, 5x, 10x), never as tokens or messages, so the right plan depends on how often you hit a cap. The chat, Cowork and Design features these plans also carry are in my full Claude pricing guide.
Claude Code Pricing Plans Explained
Claude Code is the same tool on every paid plan. The plans differ in usage, in whether the 5-hour window applies, in how much Fable you can run and in who controls billing.
Pro: $20 a month, Claude Code with 1x usage
Pro costs $20 per month billed monthly, or $200 billed up front for a year ($17 per month). It includes Claude Code with the 1x allowance that Max and Ultra multiply, and it reaches Fable, Anthropic’s top model, only through paid usage credits. Fable 5.2 needs Claude Code v2.1.321 or later, and Medium is its default effort in Claude Code.
During the Fable 5.2 launch offer, from November 9 through November 16, 2026 at 11:59 pm PT, Pro and Team standard seats get all Fable models included up to 50% of weekly usage limits, plus one Fable limit reset for accounts on paid billing before 10:00 am PT on November 9. After that, Fable returns to usage credits.
For a developer who codes in short sessions a few days a week, that allowance is enough. The trigger to leave is repeated limit hits: Anthropic’s help center tells Pro users to consider Max “if you consistently hit limits and need more capacity for larger repositories.”
Max: $100 a month for 5x Pro usage
Max costs $100 per month, billed monthly only, for 5x Pro usage every 5 hours and every week, plus Fable up to half your weekly limit and priority access at high traffic times.
At five times Pro’s price for five times Pro’s capacity, Max does not make usage cheaper; it buys longer sessions before a wall. Its $100 a month in API credits can’t be used in Claude Code or the Claude apps, including for usage credits.
Pick Max when Pro’s 5-hour window stops you on most working days. Skip it if you hit Pro’s limit only a few times a month: at the $13 average for an active day, two extra days on usage credits cost less than the $100 – $20 = $80 step up.
Ultra: $200 a month with no 5-hour limit
Ultra costs $200 per month, billed monthly only, for 10x Pro usage every week and no 5-hour limit on verified accounts. If you searched for a “20x” Claude Code plan, the $200 tier on the November 2026 pricing page is Ultra, and its multiple is 10x Pro per week.
Ultra doubles Max’s price and its weekly multiple, so the real difference is the window: Max cuts a session off every 5 hours, while a verified Ultra account lets a long agent run continue until the weekly limit. Skip Ultra if your sessions end before Max’s 5-hour cap does, since the extra $100 a month buys headroom you never reach.
Team: Standard and Premium seats
Every Team seat includes Claude Code. Standard seats cost $20 per member per month, billed annually, or $25 per member per month, billed monthly. A Standard seat gets more usage than Pro. Premium seats cost $100 per member per month, billed annually, or $125 per member per month, billed monthly. A Premium seat gets 5x a Standard seat’s usage. Team covers 2 to 150 people, and seat types can be mixed.
Each member draws on a per-seat allowance that resets on a rolling 5-hour window and a weekly window, shared with Claude chat and Cowork. That makes the seat mix your main cost lever: Premium for each developer who runs Claude Code most of the day, Standard for everyone else.
Team also brings central billing and administration, SSO, Fable for up to half of weekly limits on Premium seats, and up to $500 a month in pooled API credits that can’t be used in Claude Code.
Enterprise: $20 per seat per month plus usage
Enterprise charges $20 per seat per month, billed annually, with a 20-seat minimum. Usage is billed on top of the seat fee, at API rates. A usage-based Enterprise plan, self-serve included, has no per-seat usage limits: Claude Code use is metered consumption.
The seat floor alone is $20 x 20 = $400 a month before usage, so Enterprise buys governance, not a cheaper seat: SCIM, audit logs, role-based access, a Compliance API and a HIPAA-ready offering, each marked “No” in the Team column. The sales-assisted version is Contact sales and lists trials and invoicing on net terms. Ten Standard Team seats cost $20 x 10 = $200 a month, half that floor.
My take
Pick Pro if you rarely meet the 5-hour cap, Max if you meet it most days, and Ultra only if the 5-hour cap itself is what stops your agent runs. For a team, Standard seats with a few Premium seats are worth it until you need SCIM, audit logs or a HIPAA-ready offering; only then is Enterprise worth the 20-seat floor.
Is Claude Code Free?
No. The Free plan’s feature table marks Claude Code “No”, so the cheapest way to run it is Pro at $20 per month billed monthly or a Console account billed per token. The pricing page states Trials only in its Team and Enterprise table, where sales-assisted Enterprise shows “Yes” and Team shows “No”.
To try Claude Code before committing, buy one month of Pro, or use a Console account where you pay only for the tokens a test uses. My walkthrough on how to use Claude Code covers the install command and a first session. Once on a paid plan, my copy-paste Claude prompts for coding give that first month real tasks to run.
How Claude Code Usage Limits Work
Claude Code shares one usage pool with Claude on the web, desktop and mobile apps, and that pool has two caps on paid plans: a 5-hour session limit that starts with your first message and resets in full five hours later, and a weekly limit that resets at a fixed time assigned to your account. Anthropic publishes no fixed message or token count for any plan.
The pool drains faster with longer and more complex sessions, larger files, the model you choose, a higher effort level and connected tools, which the help center calls token-intensive. Claude Code in VS Code, Cursor and JetBrains counts toward the same limits as the terminal, and Max and Ultra add a separate weekly limit for Fable.
On Team, each seat has its own allowance; on a usage-based Enterprise plan there is no per-seat limit, only metered consumption. At a limit you can wait for the reset, move up a plan, turn on usage credits at standard API rates, or switch to a Console account for an intensive coding sprint. Inside Claude Code, /usage shows what counts against your plan and /status your remaining allocation.
Which cap you meet first decides the upgrade. If the 5-hour window cuts you off mid-afternoon, Max’s 5x every 5 hours fixes it; if a long agent run stalls on the window itself, only Ultra removes it. Chat-side limits and length limits are covered in my guide to how Claude’s usage limits work.
Claude Code API Pricing: Pay-As-You-Go Token Rates
With a Claude Console account, Claude Code consumes API tokens at standard API pricing. Opus 5.5 costs $4 per million input tokens and $20 per million output tokens, Sonnet 5.5 costs $2 and $10, Haiku 5.5 costs $0.10 and $0.50 for prompts up to 100,000 tokens, and Fable 5.2 costs $8 and $40. Fable 5.1 stays listed at $10 and $50.
| Model | Input | Output | Cache read | Notes |
|---|---|---|---|---|
| Fable 5.2 | $8 | $40 | $0.20 | Cache hit costs 2.5% of the input price; fast mode $16 input, $80 output |
| Opus 5.5 | $4 | $20 | $0.20 | Fast mode: $8 input, $40 output |
| Sonnet 5.5 | $2 | $10 | $0.10 | Cache hit costs 5% of the input price |
| Haiku 5.5, prompts up to 100,000 tokens | $0.10 | $0.50 | $0.01 | Lowest rates in this table |
| Haiku 5.5, prompts over 100,000 tokens | $0.50 | $2.50 | $0.05 | Whole request billed at the higher rate |
Sources: Claude API pricing docs and the Claude Code fast mode docs, checked November 2026.
Cached reads matter more for coding than for chat, because each Claude Code turn carries file contents, tool calls and multi-step reasoning. A session that rereads the same files pays the cache-read rate for them after the first pass, a twentieth of the Opus 5.5 or Sonnet 5.5 input price.
US-only inference multiplies every token price by 1.1x. A Console organization can set workspace spend limits on total Claude Code spend, and its rate limits apply to the organization, not to each user. On Amazon Bedrock, Google Cloud or Microsoft Foundry, Claude Code is billed per token to your cloud account, outside this page.
Take an illustrative session that sends 3 million input tokens and gets back 200,000 output tokens on Sonnet 5.5. Input costs 3 x $2 = $6 and output costs 0.2 x $10 = $2, so the session costs $8 at list price before caching.
The same session on Opus 5.5 costs 3 x $4 = $12 for input plus 0.2 x $20 = $4 for output, or $16.
API Pay-As-You-Go vs a Monthly Subscription
At Anthropic’s published average of about $13 per developer per active day, API billing costs more than Pro from 2 active days a month, more than Max from 8 and more than Ultra from 16. A developer who codes most working days pays less on a plan; one who codes on fewer than 2 days a month pays less on tokens.
| Plan | Monthly price | Active days to pass it | Math |
|---|---|---|---|
| Pro | $20 per month | 2 | 2 x $13 = $26 |
| Max | $100 per month | 8 | 8 x $13 = $104 |
| Ultra | $200 per month | 16 | 16 x $13 = $208 |
Sources: Claude Code cost docs and Anthropic’s Claude pricing page, checked November 2026.
The $13 is an average across enterprise deployments, not a price. The same cost docs put spend at $150-250 per developer per month, with 90% of users below $30 per active day. A full-time developer in that range spends less on Max at $100, as long as Max’s caps hold the workload.
Per-developer cost varies widely with usage patterns such as running multiple instances or automation. In plan mode, an agent team uses about 7x the tokens of a standard session, because every teammate holds a separate context window.
My take
The break-even decides plan versus tokens, not which plan. Stay on the Console if you code on fewer than 2 days a month; from 2 active days up, a plan costs less, so start on Pro at $20. Move to Max only when Pro’s 5-hour cap stops you on most working days. Put CI jobs on a Console workspace with a spend limit, because a plan’s 5-hour window can stop an unattended pipeline.
If OpenAI’s agent is in the running too, my breakdown of how Claude Code compares with Codex sets the two billing models side by side.
Monthly vs Annual Claude Code Pricing
Only Pro and Team bill annually: a year of Pro costs $20 x 12 = $240 billed monthly against $200 billed up front, a $40 saving. Max and Ultra bill monthly only, and Enterprise bills annually only.
| Plan | Monthly billing | Annual billing | A year billed monthly | A year billed annually | Saving |
|---|---|---|---|---|---|
| Pro | $20 per month | $200 billed up front | $20 x 12 = $240 | $200 | $240 – $200 = $40 |
| Team Standard seat | $25 per seat per month | $20 per seat per month | $25 x 12 = $300 | $20 x 12 = $240 | $300 – $240 = $60 |
| Team Premium seat | $125 per seat per month | $100 per seat per month | $125 x 12 = $1,500 | $100 x 12 = $1,200 | $1,500 – $1,200 = $300 |
Sources: Anthropic’s Claude pricing page, checked November 2026.
Annual Pro is a bet that you keep Claude Code most of the year: its one charge equals 10 months of monthly Pro ($20 x 10 = $200), so it wins only past month 10. Switching from monthly to annual credits your remaining time toward the first annual invoice, so you do not pay twice for the same days.
Costs Beyond the Subscription
Eight charges can sit outside the plan price, one per row below: usage credits at standard API rates, API charges from an ANTHROPIC_API_KEY in your environment, Console auto-reload, fast mode on Fable 5.2 or Opus 5.5, Fable on Pro, US-only inference, Enterprise usage and tax. The monthly API credits on Max, Ultra and Team offset none of them inside Claude Code.
| Cost type | When it applies | Pricing basis | How to control it |
|---|---|---|---|
| Usage credits | After a 5-hour or weekly limit on a paid plan | Standard API rates | Monthly spend limit; on Team and Enterprise, limits per org, group or member |
| ANTHROPIC_ | Claude Code signs in with the key instead of your plan | API usage charges | See Anthropic’s help article on managing API key environment variables in Claude Code |
| Console auto-reload | You chose API credits with auto-reload on | Credits added when the balance runs low | Turn it off in Console billing settings |
| Fast mode on Fable 5.2 or Opus 5.5 | You turn it on with /fast | $16 input, $80 output per million tokens on Fable 5.2; $8 input, $40 output on Opus 5.5; on subscription plans, usage credits only | Stay on standard speed |
| Fable on Pro | Any Fable use on Pro | Paid with usage credits | Use Opus or Sonnet |
| US-only inference | US inference chosen on the API | 1.1x token prices | Keep default global routing |
| Enterprise usage | All usage on usage-based Enterprise | API rates | Admin-set user and org spend limits |
| Tax | Every paid plan | Prices shown exclude applicable tax | Not applicable |
Sources: Anthropic’s Claude pricing page, the help center article on Claude Code with Pro, Max or Ultra, the Claude Code product page, the Claude API pricing docs and the Claude Code fast mode docs, checked November 2026.
The API key row bills subscribers by surprise: with ANTHROPIC_API_KEY set, Claude Code uses that key instead of your plan, so the session lands on your API bill. To keep spend inside the plan, decline the API credit option when Claude Code offers it, or log out and back in with only your plan credentials so the prompt never appears.
How Much Does Claude Code Cost for Your Team?
A year of Claude Code runs from $20 x 12 = $240 for one developer on Pro billed monthly to $100 x 10 x 12 = $12,000 for ten developers on Team Premium seats billed annually. Twenty-five developers on Enterprise pay $20 x 25 x 12 = $6,000 a year in seat fees, and at Anthropic’s average of $150-250 per developer per month, usage adds 25 x $150 x 12 = $45,000 to 25 x $250 x 12 = $75,000 a year on top.
Claude Code cost calculator
This calculator needs JavaScript. The scenario table below gives the same estimate for typical workloads.
Enter what you need. The calculator prices it on every plan at list prices and names the cheapest plan that covers it.
- List prices in USD from Claude Code’s official pricing pages, checked Nov 7, 2026. Tax and regional prices are not included.
- 12-month figures are 12 times the monthly cost at the billing you choose.
- A planning estimate at list prices, not a quote.
Enter your number of developers, the billing you want and what you need (5x usage, no 5-hour limit, central administration, SCIM), and the calculator shows every plan side by side with its monthly and 12-month cost and the cheapest plan that fits. It prices Pro month to month, because annual Pro is one $200 charge. It leaves out tax, usage credits, fast mode and Enterprise usage.
| Team size | Plan | Billing | Monthly equivalent | Annual total | Assumptions |
|---|---|---|---|---|---|
| Part-time solo developer | Pro | Monthly | $20 x 1 = $20 | $20 x 12 = $240 | 1x usage covers the work; no usage credits |
| All-day solo developer | Max | Monthly | $100 x 1 = $100 | $100 x 12 = $1,200 | Needs 5x Pro usage; no usage credits |
| Solo developer running long tasks | Ultra | Monthly | $200 x 1 = $200 | $200 x 12 = $2,400 | Verified account, no 5-hour limit |
| Team of 5 developers | Team Standard | Annual | $20 x 5 = $100 | $20 x 5 x 12 = $1,200 | Central billing; Standard seats only |
| Team of 10 heavy coders | Team Premium | Annual | $100 x 10 = $1,000 | $100 x 10 x 12 = $12,000 | Premium seats for all ten |
| Team of 25 developers needing SCIM | Enterprise | Annual | $20 x 25 = $500 | $20 x 25 x 12 = $6,000 | Seat fees only; usage at API rates adds 25 x $150 x 12 = $45,000 to 25 x $250 x 12 = $75,000 a year at Anthropic’s $150-250 per developer per month average |
Sources: Anthropic’s Claude pricing page, checked November 2026. Totals exclude tax.
I would expect most teams to mix Standard and Premium seats. Five developers with three Standard and two Premium seats on annual billing pay 3 x $20 x 12 = $720 plus 2 x $100 x 12 = $2,400, or $3,120 a year, so the seat mix matters more than the plan name.
The Enterprise seat total is only a floor, so pilot a small group with /usage before you budget the rest.
Which Claude Code Plan Should You Choose?
Pro fits part-time coding, Max all-day solo work, Ultra the long runs the 5-hour window stops, Team a dev team, Enterprise 20 or more developers who need SCIM, and the Console API scripts and occasional use.
| Your needs | Plan to consider | Why | Main limitation |
|---|---|---|---|
| You code a few hours a week and rarely hit a limit | Stay on Pro | $20 a month covers Claude Code at 1x usage | Fable only through usage credits |
| Pro’s 5-hour limit stops you on most working days | Max | 5x Pro usage every 5 hours and every week | The 5-hour window still applies |
| Long agent runs stall at the 5-hour window | Ultra | No 5-hour limit for verified accounts, 10x Pro weekly | Monthly billing only |
| You run a dev team of 2 to 150 and want one bill | Team Standard, Premium for heavy coders | Central billing and mixed seat types | Pooled API credits can’t pay for Claude Code |
| You have 20 or more developers and need SCIM or audit logs | Enterprise | SCIM, audit logs and a HIPAA-ready offering | 20-seat minimum plus usage at API rates |
| You code only occasionally (see the break-even above) | Console API | You pay only for the tokens you use | Every token is metered; you can set a workspace spend limit in the Console |
Sources: Anthropic’s Claude pricing page and the Claude Code cost docs, checked November 2026.
My take
I would start nearly every solo developer on Pro billed monthly, because a month of /usage shows which limit you actually hit. Move to Max only after the 5-hour cap stops you on most working days, and switch Pro to annual once you know you will keep it past month 10.
If no plan’s limits fit how you work, my list of Claude alternatives is sorted by the reason people leave.
Is Claude Code Worth the Price?
Claude Code is worth it for a developer who codes several days a week: at Anthropic’s average of about $13 per active day on API billing, Pro’s $20 a month costs less than 2 days of tokens. Max is worth it only when Pro’s 5-hour window stops you on most working days, Ultra only when long agent runs stall at the 5-hour window, and a developer who codes on fewer than 2 days a month pays less on the Console.
Coding drains a plan faster than chat: Anthropic advises budgeting more for a coding seat than a chat seat, since one debugging session can consume more than a day of chat.
- Skip it if:
- you want free coding help (Claude Code isn’t on Free), or you code on fewer than 2 days a month (tokens cost less).
My take
Pro at $20 is worth it for anyone who codes most days, and it is the plan I would buy first. Max is worth it when Pro’s 5-hour window stops you on most working days, not a few times a month.
For a verdict on Claude itself rather than its price, see my Claude review.
To compare tools beyond Anthropic’s, my ranking of other AI coding tools by job and budget widens the field.
Billing, Discounts and Cancellation
The discount routes are annual billing on Pro and Team, volume and committed-spend discounts on Enterprise and the API through Anthropic’s sales team, at rates it does not publish, and an Education plan for universities. You can cancel anytime, the plan stays active until the billing period ends, and canceling at least 24 hours before renewal avoids the next charge.
Payments are generally non-refundable, except where Anthropic’s Consumer Terms of Service say otherwise or local law requires it, as its pricing page FAQ states. Upgrades take effect right away with unused time credited; downgrades take effect at the end of the billing period, and a Pro, Max or Ultra account moves to Free once the paid period ends.
How I Verified Claude Code Pricing
My sources were Anthropic’s pricing page as served to a US visitor, its API pricing docs, the Claude Code cost docs and its help-center articles on Claude Code with Pro, Max or Ultra, on Claude Code with Team or Enterprise, and on usage and length limits, all read on November 7, 2026, in US dollars. A second copy of the pricing page fetched from outside the US showed the same dollar amounts.
Every total is list price x seats x months. The break-even uses Anthropic’s $13 average per developer per active day, an estimate from enterprise deployments rather than a price. Anthropic does not publish a token or message allowance for any plan, an annual price for Max or Ultra, or its discount rates.
Frequently Asked Questions
Is there a Claude Code student discount?
The pricing page lists no individual student price for Claude Code. Its Education plan is a university-wide plan an institution buys, giving students, faculty and staff access at discounted rates, so a student should first ask whether their university has it. My list of the best AI tools for students covers other study tools.
Does Claude Code in VS Code or JetBrains cost extra?
No. A Pro, Max or Ultra plan covers Claude Code in VS Code, Cursor and other VS Code forks, and JetBrains IDEs such as IntelliJ and PyCharm, and a Team or Enterprise seat covers them too. If you are weighing Cursor’s own plans instead, my Cursor vs Claude Code comparison sets the two bills side by side. On Team and Enterprise, IDE usage is limited and billed the same way as terminal usage.
Can I use one Claude plan for chat and Claude Code?
Yes. One Pro, Max or Ultra subscription covers the Claude web, desktop and mobile apps and Claude Code in your terminal, and Team and Enterprise seats work the same way. Because both draw from one pool, a long morning chat leaves less room for coding that afternoon.
Final Pricing Verdict
Final recommendation
Start on Pro at $20 per month unless you already code all day. Move to Max when Pro’s 5-hour limit stops you on most working days, and to Ultra only when the 5-hour window itself is the blocker.
Dev teams should buy Team seats, Standard for most developers and Premium for the heavy ones, and leave Enterprise until 20 or more developers need SCIM or audit logs. Occasional or scripted use belongs on the Console API with a workspace spend limit.
Whichever plan you pick, check /usage after the first two weeks: it shows how much of each limit you used.
