Claude has a free plan and paid plans:
- Pro at $17/month (billed annually, $20 monthly) for regular personal use;
- Max from $100/month for heavy users who need much higher limits;
- Team at $20 per seat (billed annually, $25 monthly) and Enterprise at $20 per seat plus usage for organizations
Claude Code is included in every paid plan (Pro, Max, Team and Enterprise). Developers using the API pay per million tokens, from $0.10 input / $0.50 output (Haiku 5.5) to $10 / $50 (Fable 5.1) depending on the model.
Below: every plan’s price and limits, what Claude Code really costs, the full API price table, and five ways to pay less for the same work. Prices last verified October 27, 2026.
Key terms used in this guide: token, prompt caching, and context window.
Claude pricing at a glance
Claude AI pricing has two parts: five subscription plans that share one usage pool per account, and a separate pay-per-token API. The table shows what each plan costs in US dollars and how much room it gives you compared with Free.
| Plan | Price/month | Who it’s for | Claude Code included? | Usage vs Free |
|---|---|---|---|---|
| Free | $0 | Trying Claude, occasional questions | No | Baseline, resets every 5 hours |
| Pro | $20, or $17 billed annually (200 up front) | Individuals who use Claude most days | Yes | At least 5x Free per session |
| Max 5x | $100, monthly only | Heavy daily users | Yes | 5x Pro per session |
| Max 20x | $200, monthly only | All-day work, parallel coding sessions | Yes | 20x Pro per session |
| Team, Standard seat | $25 per seat, or $20 billed annually | Teams of 2 to 150 | Yes | More than Pro |
| Team, Premium seat | $125 per seat, or $100 billed annually | Team members who code or work in Claude all day | Yes | 5x a Standard seat |
| Enterprise | $20 per seat billed annually (20-seat minimum), plus usage at API rates | Large organizations with security and compliance needs | Yes | Pay-as-you-go at API rates |
Prices last verified: October 27, 2026, on claude.com/pricing and the Claude Help Center. Prices don’t include tax, and local or in-app prices can differ.
Two rules explain most of the gaps between Claude plans. Every plan has a five-hour session limit, and paid plans add a weekly limit on top. Chat, the desktop app, mobile and Claude Code all draw from the same pool, so a long coding session leaves less room for chat later that day.
The Free plan: what it covers
Free covers chat on web, desktop and mobile, web search, file creation, memory, connectors and up to five projects, with the Sonnet and Haiku models. Opus, Fable, Research and Claude Code need a paid plan, and limits reset every five hours. For how those resets and caps work, see our Claude usage limits guide.
Claude Pro: $20 a month for everyday use
The Claude Pro price is $20 a month billed monthly, or $17 a month on an annual plan paid as $200 up front. Annual billing saves $40 a year, and Pro is the only individual plan that offers it.
Pro gives you at least five times the Free plan’s usage per session, plus Claude Code, Opus models, Research, Projects, Claude in Chrome, Claude for Microsoft 365 and scheduled tasks. Fable is available on Pro through usage credits, which bill at standard API rates once you switch them on.
Short verdict: Pro fits anyone who works with Claude most days and usually finishes before the limit does. If you run out several times a week, Max is the next step. For the full value math, read is Claude Pro worth it.
Claude Max: two tiers for heavy users
The Claude Max plan comes in two sizes: Max 5x at $100 a month and Max 20x at $200 a month, both billed monthly. You get everything in Pro, with five or twenty times Pro’s usage per five-hour session, higher output limits, priority access at busy times and early access to new features. Max also includes monthly API credits for building on the Claude Platform: $100 on Max 5x and $200 on Max 20x. Fable models are part of the plan, capped at 50% of your weekly limits.
When does it pay off? Extra usage on Pro bills at API rates, so if you regularly spend more than about $80 a month in usage credits on top of Pro, Max 5x costs less for the same work. Max 20x makes sense when you run several Claude Code sessions in parallel or keep agents working for most of the day.
Claude Team and Enterprise: per-seat pricing
The Claude Team plan is for groups of 2 to 150 people. Standard seats cost $25 per seat per month, or $20 billed annually, and give more usage than Pro. Premium seats cost $125, or $100 billed annually, with five times the usage of a Standard seat. You can mix the two, so only your heaviest users pay for Premium.
Every Team seat includes Claude Code, Cowork, Design, Slides and Docs, plus the admin layer – single sign-on, central billing, enterprise search across your company’s tools, connector controls and no model training on your content by default. The plan also includes up to $500 a month in pooled API credits. Our guide to Claude for teams covers setup and day-to-day features.
Enterprise costs $20 per seat per month, billed annually with a 20-seat minimum, plus usage billed at API rates. It adds SCIM, audit logs, a Compliance API, custom data retention, role-based access, IP allowlisting, spend limits per user and per organization, and a HIPAA-ready option. You can set it up self-serve or go through sales for invoicing, trials and committed-spend discounts. Our Claude Enterprise breakdown explains what those controls do in practice.
Claude Code pricing: plans, usage and a weekly cost example
There’s no separate Claude Code subscription. You pay for it in one of two ways: through a paid Claude plan, where it shares your usage limits, or through the API, where every token is billed.
| Plan | Included? | How usage works | Good for |
|---|---|---|---|
| Free | No | Not available | Chat only |
| Pro | Yes | Shares the 5-hour and weekly limits with chat; usage credits at API rates when you run out | Short daily sessions, side projects |
| Max 5x / 20x | Yes | Same shared pool at 5x or 20x Pro, plus $100 or $200 in monthly API credits | Multi-hour daily coding; parallel sessions on 20x |
| Team | Yes, every seat | Standard seats get more than Pro, Premium seats 5x Standard; pooled API credits | Engineering teams of 2 to 150 |
| Enterprise | Yes | Seat fee plus usage at API rates, with admin spend limits | Large teams that need cost controls and audit logs |
| API (Console) | Pay per token | Billed at the rates of the model you choose, no plan limits | Automation, CI pipelines, occasional use |
Prices last verified: October 27, 2026.
On a subscription, Claude Code counts against the same session and weekly limits as chat, and it uses them faster. The agent reads files, runs commands and resends context on every turn, so a large repository or a long session with a strong model burns through the allowance sooner than a conversation would. You can check what’s left in Settings at any time. If you’re just starting, our guide on how to use Claude Code for beginners shows how to keep sessions lean.
API billing suits work that doesn’t fit a personal plan: scripts and CI jobs that run Claude Code without a person at the keyboard, shared service accounts, or a developer who only codes with it a couple of times a month. The Console also gives you cost tracking by project and hard spend limits. For steady daily use by one person, a subscription usually costs less.
A typical week of coding
Anthropic’s Claude Code documentation puts the average across enterprise API deployments at about $13 per developer per active day, or $150 to $250 per developer per month, with 90% of users staying under $30 per active day. A developer who codes five days a week at that average spends roughly $65 a week on tokens. The same week costs about $4.60 on Pro (20 a month spread over 4.3 weeks) or about $23 on Max 5x, as long as the work fits inside the plan’s limits.
That gap is why most individual developers start on Pro, move to Max when limits get in the way, and keep API billing for automation. If neither option fits your budget or workflow, we compared other tools in Claude Code alternatives.
Claude API pricing for every current model
The API bills per million tokens, and output costs five times more than input on every current model. If you’re not sure what counts as a token, here’s what a token in AI is. One detail affects estimates: Claude 4.7 and later models use a newer tokenizer that produces about 30% more tokens for the same text than older models, so count tokens on the model you plan to use.
| Model | Input / 1M | Output / 1M | Cached input (cache hit) | Batch (input / output) |
|---|---|---|---|---|
| Fable 5.1 | $10 | $50 | $0.25 | $5 / $25 |
| Opus 5.5 | $4 | $20 | $0.20 | $2 / $10 |
| Sonnet 5.5 | $2 | $10 | $0.10 | $1 / $5 |
| Haiku 5.5, prompts up to 100K tokens | $0.10 | $0.50 | $0.01 | $0.05 / $0.25 |
| Haiku 5.5, prompts over 100K tokens | $0.50 | $2.50 | $0.05 | $0.25 / $1.25 |
Prices last verified: October 27, 2026, on Anthropic’s API pricing documentation and claude.com/pricing. All prices in USD.
Writing to the cache costs 1.25 times the input price for a five-minute cache or twice the input price for a one-hour cache. Claude Mythos 5.1 is billed at the same rates as Fable 5.1, but access is limited to a small group of organizations, so most developers will choose from the four models above. That choice is about capability as much as price, and our guide to the best Claude model matches each one to the tasks it handles well.
A few other charges can show up on the bill. US-only inference costs 1.1 times standard rates. Fast mode on Opus 5.5 runs up to 2.5 times faster for $8 input and $40 output per million tokens, and it can’t be combined with the Batch API. Web search costs $10 per 1,000 searches plus the tokens it adds. Every current model except Haiku 5.5 includes the full 1M-token context window at standard rates, so a 900,000-token request costs the same per token as a short one.
Worked example: summarizing 1,000 support emails a day
Say each email is about 800 tokens, you reuse a 1,500-token instruction prompt that stays cached, and each summary runs about 150 tokens. A day’s run comes to 0.8 million fresh input tokens, 1.5 million cached tokens and 0.15 million output tokens.
| Setup | Cost per day | Cost per 30 days |
|---|---|---|
| Haiku 5.5 | about $0.17 | about $5 |
| Sonnet 5.5 with the Batch API | about $1.63 | about $49 |
| Sonnet 5.5 | about $3.25 | about $98 |
| Opus 5.5 | about $6.50 | about $195 |
Cache writes add a few cents on top. On Sonnet 5.5, output is $1.50 of the $3.25 daily bill, so asking for tighter summaries saves more than trimming the instructions. For straightforward triage like this, Haiku 5.5 handles the volume for about $5 a month.
Prompt caching, the Batch API and 5 ways to pay less
Most savings come from five habits. Before them, here’s how the numbers look for one person on a subscription.
Worked example: one person, one month
A freelance marketer uses Claude for about two hours most workdays for drafting, research and the occasional spreadsheet. On Pro that’s $20 a month, or $200 a year with annual billing instead of $240. In a busy launch month she hits the five-hour limit four times and switches on usage credits to keep going, which adds around $15 at API rates. Her month costs $35. If those extra credits start passing $80 a month, Max 5x at $100 becomes the cheaper plan.
1. Cache the parts of your prompt that repeat
Prompt caching stores a reused prefix, such as a system prompt, a style guide or codebase context. On Opus 5.5 and Sonnet 5.5, a cache hit costs 5% of the normal input price; on Fable 5.1 it’s 2.5%. A five-minute cache pays for itself after one reuse, a one-hour cache after two.
2. Send non-urgent work through the Batch API
Batch requests cost half the standard price for input and output on every model. Results come back asynchronously, so the Batch API fits overnight jobs like classification, summaries, translations and evaluations. Batch and caching discounts stack.
3. Route each task to the cheapest model that handles it
Start with Haiku 5.5 for extraction, tagging and routing, use Sonnet 5.5 for most production work, and send only the hard cases to Opus 5.5 or Fable 5.1. Fable 5.1 output costs 100 times as much as Haiku 5.5 output on short prompts, so routing often saves more than any single discount.
4. Keep context and output lean
Every token you send or receive is billed, and in long chats or coding sessions the whole history goes back to the model on each turn. Start a fresh conversation for a new task, attach only the files the model needs, and ask for the length you want. On a subscription, the same habits stretch your limits further.
5. Match the billing model to how you work
Steady personal use is cheapest on a plan, while spiky or automated use is cheapest on the API. Pro’s annual billing saves $40 a year, teams can mix Standard and Premium seats, and verified nonprofits get up to 75% off Team and Enterprise. Max and Team also include monthly API credits, so use those before topping up a separate Console balance. Getting the billing model right is often the quickest way to lower your Claude cost.
Claude vs ChatGPT vs Gemini pricing
At around $20 a month the three are close. The differences show up at the low end, at the top end and in API prices.
| Claude | ChatGPT | Gemini | |
|---|---|---|---|
| Budget plan | None | Go, $8 | Google AI Plus, $4.99 |
| Standard plan | Pro, $20 | Plus, $20 | Google AI Pro, $19.99 |
| Top consumer plan | Max, $100 or $200 | Pro, $100, $200 or $500 | Google AI Ultra, $99.99 or $199.99 |
| Flagship API model (input / output per 1M) | Fable 5.1, $10 / $50 | GPT-6 Astra, $10 / $50 | Gemini 4 Argon, $2 / $10 introductory rate through November 30 |
Prices last verified: October 27, 2026. US list prices per month.
Claude has no plan under $20, but it’s the only one of the three with an annual discount on its standard plan. On the API, Fable 5.1 and GPT-6 Astra cost the same per token, and Gemini 4 Argon is cheaper while Google’s introductory rate lasts. For a closer look at the $20 tier, read ChatGPT Plus vs Claude Pro; for differences beyond price, see Claude vs ChatGPT.
Which Claude plan should you choose?
| Who you are | Plan | Why |
|---|---|---|
| Trying Claude or asking a few questions a week | Free | Sonnet and Haiku cover everyday chat at no cost |
| Using Claude for work most days | Pro | Claude Code, Opus and Research for $20, or $17 a month billed annually |
| Hitting Pro limits several times a week | Max 5x | Five times Pro’s usage, plus $100 in monthly API credits |
| Coding or running agents most of the day | Max 20x | Room for parallel sessions, plus $200 in monthly API credits |
| A team of 2 to 150 | Team, mixing seat types | Shared admin and billing; Premium seats only for heavy users |
| An organization that needs SCIM, audit logs or HIPAA readiness | Enterprise | Compliance features, spend controls and usage at API rates |
| A developer building an app or automation | API | Pay only for tokens, with batch and caching discounts |
| A verified nonprofit | Team or Enterprise with the nonprofit discount | Up to 75% off seat prices |
If you haven’t used Claude much yet, our guide on how to use Claude AI for beginners helps you get real work out of Free before you pay for anything.
Whatever plan you pick, you’re paying for results. Learning to brief Claude well means fewer messages do more work, on any plan. Coursiv’s 28-Day AI Certificate Program builds the prompting habits that make each message count, so your limits and token budget go further.
FAQ
How much does Claude cost per month?
How much does Claude Code cost?
What’s the difference between Claude Pro and Max?
How much does the Claude API cost per token?
Is Claude cheaper than ChatGPT?
Are there student, education or nonprofit discounts?
Can I cancel or downgrade a Claude plan at any time?
Yes. Cancel in Settings > Billing at least 24 hours before renewal, and your plan stays active until the end of the billing period. Pro and Max accounts then move to Free, and your chats, projects and files stay. Subscriptions bought through the App Store or Google Play are cancelled there, and payments are generally non-refundable except where Anthropic’s terms or local law say otherwise.
Want your plan to go further? Coursiv’s AI Certificate Program gives you a structured way to build everyday AI skills, with a certificate of completion at the end.