Claude pricing in 2026: Free, Pro, and Team plans explained
Read on October 26, 2026. Every Claude plan, price and usage multiple below was read that day from Anthropic's pricing page, its Pro, Max, Team and Fable help articles, and its model and API pricing documentation (sources and method at the end). This is a read of what Anthropic publishes, not a hands-on test of any plan. Plan prices for ChatGPT and Google were read the same day from their own pricing pages; the few comparison cells those reads do not cover are earlier figures and are shown in italics.
Anthropic sells Claude across five tiers: Free ($0), Pro ($20/month), Max ($100 or $200/month), Team (from $20/seat/month) and Enterprise ($20/seat/month plus usage), with a separate per-token API for developers. Most individuals should start on Free and move to Pro only after repeatedly hitting the usage limit, or the day they need Claude Code, Research or the Opus models. Max is for people who run out of Pro; Team is for two or more people who need shared projects and admin controls; Enterprise is for the security and compliance controls Team does not carry. This guide lays out every plan against the others, shows what a unit of usage costs on each, compares the price to ChatGPT and Gemini, and ends with a decision tree and a verdict for each kind of reader.
Entry paid plan
Pro, billed monthly. $17/mo on annual billing, $200 up front.
at least 5x FreeCost per Pro-sized session
Pro, Team Standard, Max 5x and Team Premium, monthly billing.
flat from 1x to 6.25xThe only volume discount
$200 for 20x Pro. Annual billing takes Team seats to $16.00.
half price per sessionWhat are all the Claude plans at a glance?
| Plan | Price | Models | What it adds |
|---|---|---|---|
| Free | $0read 2026-10-26 | Sonnet, Haiku | Chat, web search, file creation and code execution, memory, connectors, Artifacts, Claude Design, Slides and Docs; Projects up to 5 |
| Pro | $20/moread 2026-10-26 ($17/mo annual, $200 up front) | + Opus; Fable on usage credits | At least 5x Free usage per session, Claude Code, Research, Claude in Chrome, Microsoft 365, Dashboards (beta), Science, hand off and schedule tasks |
| Max 5x | $100/moread 2026-10-26, monthly billing only | + Fable within 50% of weekly limits | 5x Pro usage, higher output limits, priority access at high traffic times, $100/mo API credits |
| Max 20x | $200/moread 2026-10-26, monthly billing only | Same as Max 5x | 20x Pro usage, $200/mo API credits |
| Team Standard | $20/seat/moread 2026-10-26 annual ($25 monthly), 2 to 150 seats | + Opus; Fable on usage credits | 1.25x Pro per-session usage, Claude Code, shared projects, SSO, central billing, up to $500/mo pooled API credits |
| Team Premium | $100/seat/moread 2026-10-26 annual ($125 monthly) | + Fable within 50% of weekly limits | 5x Standard seat usage (6.25x Pro) |
| Enterprise | $20/seat/moread 2026-10-26 annual, 20-seat minimum, plus usage at API rates | All models | Everything in Team plus SCIM, audit logs, role-based access, custom data retention, HIPAA-ready offering, Claude Security (beta) |
Swipe the table sideways on a phone.
Which plan opens which capability
22 rows where the five tiers differ, grouped by where the dots stop. Read 2026-10-26.
The same matrix as a text table (22 rows)
| Capability | Free | Pro | Max | Team | Enterprise |
|---|---|---|---|---|---|
| Projects | Up to 5 | Yes | Yes | Yes | Yes |
| Claude Opus | No | Yes | Yes | Yes | Yes |
| Research | No | Yes | Yes | Yes | Yes |
| Claude Code | No | Yes | Yes | Yes | Yes |
| Claude in Chrome | No | Yes | Yes | Yes | Yes |
| Claude for Microsoft 365 | No | Yes | Yes | Yes | Yes |
| Claude Dashboards (beta) | No | Yes | Yes | Yes | Yes |
| Claude Science | No | Yes | Yes | Yes | Yes |
| Claude Fable | No | Usage credits | 50% of weekly limits | 50% of weekly limits on Premium seats | Yes |
| Priority access at high traffic times | No | No (Pro help article says yes) | Yes | Yes | Yes |
| Monthly API credits | No | No | $100 (5x) or $200 (20x) | Up to $500 | No |
| Project sharing and collaboration | No | No | No | Yes | Yes |
| Single sign-on (SSO) | No | No | No | Yes | Yes |
| Enterprise search | No | No | No | Yes | Yes |
| Claude Motion (beta) | No | No | No | Yes | Yes |
| Role-based access | No | No | No | No | Yes |
| SCIM | No | No | No | No | Yes |
| Audit logs | No | No | No | No | Yes |
| Compliance API | No | No | No | No | Yes |
| HIPAA-ready offering | No | No | No | No | Yes |
| Custom data retention controls | No | No | No | No | Yes |
| Claude Security (beta) | No | No | No | No | Yes |
Rows that are the same on every plan are left out: Sonnet and Haiku, chat on web, mobile and desktop, web search, memory, files and code execution, Artifacts, connectors, skills, voice mode, desktop extensions, incognito chats, user preferences, and Claude Design, Slides and Docs are all marked Yes from Free up. The first band is the practical one for an individual: seven gates Free leaves shut plus the Projects cap, all opened by Pro. Everything below that is about volume, collaboration or compliance, and none of it is a reason for a solo user to pay more than $20.
What does each Claude plan cost per unit of usage?
Anthropic never prints a message count, but it does print multiples, and the multiples can be priced. Take Pro's per-session usage allowance as one unit, a Pro-sized session. Then Anthropic's own statements give every paid plan a size: Max is 5x or 20x Pro, Team Standard seats include 1.25x the Pro plan's per-session usage allowance, and Team Premium seats include 6.25x (5x a Standard seat, which is also what the pricing page says). Free is at most one fifth of a unit, because Pro is "at least 5x more usage per 5-hour session than Free". Divide each plan's price by its size and the ladder flattens out.
What one Pro-sized session costs
Price divided by per-session multiple of Pro. Filled bar: monthly billing. Outline bar: annual billing. Read 2026-10-26.
How much session usage each plan buys
In Pro-sized sessions, drawn to scale. Every multiple is a floor or a published ratio, not a message count.
- Free$00.2x or less
- Pro$20/mo1x
- Team Standard$25/seat/mo1.25x
- Max 5x$100/mo5x
- Team Premium$125/seat/mo6.25x
- Max 20x$200/mo20x
Two practical readings. First, there is no volume discount until Max 20x: buying Max 5x because you expect to need "a bit more" than Pro costs exactly five Pro subscriptions, so the case for it is concurrency and fewer interruptions, not savings. Second, a Team Standard seat on annual billing is 67 cents cheaper per unit than Pro on annual billing ($16.00 against $16.67 per Pro-sized session), but it needs at least two members, and Anthropic says nothing about how its weekly limit compares.
What does the Claude Free tier actually include?
Claude's free tier is a standing plan, not a trial and not a crippled demo. Anthropic's pricing page calls it "Free for everyone" and lists chat on web, desktop and mobile, web search, file creation and code execution, memory across conversations, connectors, Artifacts, and Claude Design, Slides and Docs, with Projects limited to 5. Sonnet and Haiku are on every plan. The context window is "up to 1M, varies by model" on every plan, so it does not separate Free from Pro. No credit card is listed for Free.
Claude's free tier is genuinely useful, not a crippled demo.Free tier
The catch is usage, and Anthropic is careful about it. Its pricing page says what you can do "depends on the length and complexity of your conversations, the model you choose, and the features you use, so there's no fixed message count." Free usage resets on a rolling five-hour session window. Older versions of this page, and plenty of others, quoted "10 to 15 messages" and a "4 to 8 hour reset" for Free. Neither number appears in Anthropic's pages, so we removed them. The only figure with a source behind it is relative: Pro is at least 5x Free per session. Our companion pages go deeper, with the full capability count in Is Claude free? and the mechanics in Claude free plan limits.
Best for: casual users, people trying Claude for the first time, students, and anyone who rarely sees the limit message.
For a deeper look at what’s included, including the capability-by-plan matrix that counts the 29 features Anthropic tabulates and shows which of them Free includes, see our full guide: Is Claude free? The free tier explained (2026).
What do you get with Claude Pro at $20/month?
Claude Pro is the $20/month individual subscription that opens the gates Free leaves shut. At $20/monthread 2026-10-26 (or $200/year billed up front, about $17/month), Anthropic's pricing page lists these additions to Free:
- At least 5x Free usage per session, the single biggest practical difference. Pro also has a weekly limit that applies across all models, and your activity on web, desktop, mobile and Claude Code draws from one pool
- Opus 5.5, the current Opus model, for long-running agentic coding and knowledge work. Fable 5.1, above Opus, runs on pay-as-you-go usage credits on Pro
- Claude Code, the terminal coding agent, included in every paid plan. It has no standalone price: it shares your plan's usage limits, and heavy users can switch to pay-as-you-go API credits through a Console account (see our Claude Code pricing guide)
- Research, multi-step web research, plus Claude in Chrome and Claude for Microsoft 365 and Outlook
- Claude Dashboards (beta) and Claude Science, and the ability to hand off and schedule tasks
- Projects with no cap listed: Free is "Up to 5" and Pro says Yes, so "unlimited" is our shorthand, not Anthropic's word
- Usage credits, so you can keep working past the limit at standard API rates
Memory, connectors, web search and file creation are on Free too, so they are not reasons to upgrade. Anthropic's Pro help article still lists "Priority access to Claude during high-traffic periods" as a Pro benefit while the plan table marks priority access Max and above only; the two pages disagree, so do not upgrade for it alone. We removed the older "approximately 45 messages per 5-hour window" figure: it was community-reported and Anthropic does not publish it.
Best for: developers, writers, researchers, and anyone who uses Claude as a daily work tool. If you code, Claude Code alone can justify the subscription.
If you code, Claude Code alone can justify the subscription.Pro verdict
What is Claude Max and when is it worth $100 or $200 a month?
Claude Max is the high-usage individual tier, sold at two prices, $100/month for 5x Pro usage and $200/month for 20xread 2026-10-26, monthly billing only. It exists for one reason: more usage per session. If you consistently hit Pro's ceiling inside a single five-hour session, Max removes that friction. We pressure-test the value math in Is Claude Max worth it?
Max 5x ($100/monthread 2026-10-26)
Five times the Pro usage allowance per session, plus the features Anthropic lists above Pro: higher output limits, priority access at high traffic times, early access to advanced features, and $100 a month in API credits to build on the Claude Platform. Max also includes Fable 5.1 within up to 50% of your weekly limits, where Pro reaches it only through usage credits. Anthropic frames the plan as a multiplier of Pro and never publishes a message count, so any per-session number you see elsewhere is a guess.
Max 20x ($200/monthread 2026-10-26)
Twenty times Pro, with $200 a month in API credits. A weekly limit applies across all models on Max too, so the multiple is a per-session figure, not an unlimited plan. This is the one plan with a better price per unit of usage than Pro: $10.00 per Pro-sized session against $20.00, as the figure above shows. If you know you will use the extra allowance, 20x is cheaper per unit than 5x; if you will not, it is $100 more for nothing.
Who needs Max? Very few people: a developer running Claude Code for hours at a stretch, a researcher processing long documents all day, or a content professional producing high volumes of AI-assisted work. If you are not hitting Pro's limit at least 3 or 4 times a week, Max is a waste of money.
What is Claude Team and who actually needs it?
Claude Team is the multi-seat plan for 2 to 150 seats, and it adds collaboration and administration on top of Pro's features. Standard seats are $20 per seat per month on annual billing, $25 monthlyread 2026-10-26; Premium seats are $100 annual or $125 monthly, and you can mix seat types. For seat economics and where Team gives way to Enterprise, see Claude Team vs Enterprise pricing. At the Standard tier you get:
- Everything in Pro, including Opus, Claude Code, Research and the Chrome and Microsoft 365 tools. Anthropic's help article says Standard seats include 1.25x the Pro plan's per-session usage allowance
- Premium seats give 5x a Standard seat, which is 6.25x Pro per session. On Premium seats Fable runs within 50% of weekly limits; on Standard seats it runs on usage credits
- Collaboration: project sharing, enterprise search, and Claude Motion and Dashboards (beta)
- Administration: single sign-on, central billing, spending caps at organization and user level, and an Owner role. Model training on your content is off by default
- Up to $500 a month in pooled API credits, as listed on the pricing page
Best for: small to mid-size teams that need shared context and admin oversight. Claude Code on every seat is the headline value for teams working in large codebases.
Minimum seats: two. Solo users should buy Pro. Teams past 150 seats move to Enterprise.
What does Claude Enterprise add on top of Team?
Claude Enterprise is the compliance and scale tier. Anthropic prices it at $20 per seat per month billed annually, with a 20-seat minimum, plus usage billed at API ratesread 2026-10-26, so the cost scales with the models and tasks your people run. You can start self-serve or talk to sales; Enterprise is annual-only. Everything in Team plus:
- SCIM, role-based access and audit logs, plus a Compliance API for observability
- Custom data retention controls and a HIPAA-ready offering
- Network-level access control with IP allowlisting, and admin spend limits at user and organization level
- Claude Security (beta)
- Sales-assisted extras: invoicing and net terms, AWS Marketplace availability, tiered incentives on committed spend and customer success support at spend thresholds
Best for: companies that need SCIM, audit logs or HIPAA-ready controls, or that can meet the 20-seat floor and want usage billed at API rates. Earlier versions of this page also named SOC 2 and custom deployments; we did not find either on the pages read today, so we no longer state them. Ask Anthropic's sales team for a quote.
How does the Claude API price differently from the subscription?
This trips people up: the Claude API is a separate, pay-per-token product, not part of any subscription. The $20/month Pro subscription gives you claude.ai and Claude Code; the API is metered per token on the Claude Console and designed for developers building applications. For worked cost examples and batch-discount math, see our Claude API pricing guide.
Current API pricing (read 2026-10-26)
| Model | Input (per 1M tokens) | Output (per 1M tokens) |
|---|---|---|
| Fable 5.1 (top tier) | $10.00 | $50.00 |
| Opus 5.5 (current Opus) | $4.00 | $20.00 |
| Sonnet 5.5 (current Sonnet) | $2.00 | $10.00 |
| Haiku 5.5 (prompts up to 100K tokens / over 100K) | $0.10 / $0.50 | $0.50 / $2.50 |
| Legacy: Opus 5, Opus 4.8 to 4.5 / Sonnet 4.6 and 4.5 / Haiku 4.5 | $5.00 / $3.00 / $1.00 | $25.00 / $15.00 / $5.00 |
| Retired on the Claude API (Opus 4.1 and Opus 4 / Haiku 3.5), still on some cloud platforms | $15.00 / $0.80 | $75.00 / $4.00 |
Swipe the table sideways on a phone.
Output price per million tokens, current models
Bar width is the output price divided by $50. Read 2026-10-26.
- Fable 5.1in $10.00$50.00
- Opus 5.5in $4.00$20.00
- Sonnet 5.5in $2.00$10.00
- Haiku 5.5in $0.10$0.50
Prompt caching and a 50% batch discount both exist, so the table is the list-price ceiling, not the floor. A million tokens is roughly 555,000 words on the current tokenizer, per Anthropic's model documentation (older models fit about 750,000). The figures come from Anthropic's API pricing documentation and models overview. A typical conversation of 3,000 to 5,000 tokens, with half of them output, costs about $0.02 to $0.03 at Sonnet 5.5 prices ($2 in, $10 out), and under a cent on Haiku 5.5.
When to use the API vs subscription: if you are a developer integrating Claude into your own product or running automated workflows, use the API. If you are using Claude as a personal or team productivity tool, use the subscription.
How does Claude pricing compare to ChatGPT and Gemini?
| Feature | Claude Pro ($20/mo) | ChatGPT Plus ($20/mo) | Google AI Pro ($19.99/mo) |
|---|---|---|---|
| Top model included | Opus 5.5 (Fable 5.1 on usage credits) | GPT-6 (named on the Free card, ChatGPT pricing page, read 2026-10-26) | Gemini 4 Argon, plus Gemini 3.1 Pro (Google's page, read 2026-10-26) |
| Coding agent | Claude Code, included in every paid plan | Codex (cloud, IDE, local CLI) | Not compared here |
| Image generation | Not listed on Anthropic's plan table | Yes | Image, music and video models, expanded access (Google's page) |
| Video generation | Not listed on Anthropic's plan table | Not listed on its pricing page | Video models, expanded access (Google's page) |
| Web search | Yes (also on Free) | Yes | Yes (Google Search) |
| Research mode | Research | Deep Research | Deep Research in Gemini (Google's page) |
| Context window | Up to 1M tokens, varies by model | 54K Instant / 256K reasoning | Up to 1M (per Google) |
| Connectors and integrations | Connectors (also on Free) | Custom GPTs, connectors | Gemini in Gmail, Docs and Sheets (Google's page) |
| Annual option | $17/mo ($200 up front) | Not listed on its pricing page | See Google's plan page |
| Monthly price | $20 (read 2026-10-26) | $20 (read 2026-10-26); Go is $8 | $19.99 (read 2026-10-26); Plus is $4.99 |
| Heavy-use tier | Max, from $100 | Pro, from $100 (read 2026-10-26) | Ultra, from $99.99 (read 2026-10-26) |
Swipe the table sideways on a phone.
Prices and the cells marked "read 2026-10-26" come from ChatGPT's and Google's own pricing pages that day (US). Italic cells (Codex, context window, ChatGPT image and annual rows, and similar) are earlier figures from our September 2026 read, not re-read.
The honest comparison
Claude Pro wins, in our view, for: developers (Claude Code is included in every paid plan, not sold separately), long-document analysis and writing. Opus is strong at nuanced, careful prose.
ChatGPT Plus wins for: users who want image generation, custom GPTs and the most mainstream integrations. If pricing is your deciding axis, we run the dollar-for-dollar math in Claude vs ChatGPT pricing.
Google AI Pro wins for: users deep in the Google ecosystem (Gmail, Drive, Docs, Sheets), those who value Google Search integration, and anyone who wants cloud storage bundled into the same bill (Google lists 5 TB on Pro).
All three entry plans cost about $20 a month, and all three sell a heavy-use tier from about $100 (Claude Max from $100, ChatGPT Pro from $100, Google AI Ultra from $99.99, each read October 26). The cheaper rungs differ: ChatGPT Go is $8 and Google AI Plus is $4.99, while Claude's cheapest paid plan is Pro at $20. The AI Price Tracker lines them up. None is objectively "the best": your use case decides. For detailed head-to-head analysis, see our ChatGPT vs Claude vs Gemini comparison.
Free ($0)
No credit card, email sign-up
The 1x usage baseline
Pro ($20/mo)
$17/mo annual, $200 up front
At least 5x the usage
Is $20/month worth it? The break-even arithmetic
Instead of guessing how many hours Claude saves you, flip it: how many hours must it save before $20 is paid back? That is the monthly price divided by what an hour of your time is worth.
Computed as $20 divided by the hourly rate. On annual billing ($16.67) the figures fall by about 17 percent. The saving is yours to estimate; Anthropic publishes none.
If you’re a developer
Claude Code is included in Pro. At a developer’s hourly value, it only has to save about 12 to 24 minutes a month to cover the $20, and it draws from the same usage pool as your chats. Verdict: worth it for almost any working developer.
If you’re a writer or content creator
Opus 5.5 is Anthropic’s model for knowledge work, and Pro lifts you from the free allowance to at least five times as much per session. If you write daily, the arithmetic above is easy to clear. Verdict: worth it if you write daily; the free tier is fine for occasional writing.
If you’re a researcher
Research is a Pro-and-up capability that Free does not include, so the question is whether you run multi-step research weekly. Verdict: worth it for regular researchers.
If you’re a student
The free tier handles most academic tasks. Upgrade to Pro only if you code regularly (for Claude Code) or consistently hit the free limit. $20 a month on a student budget is meaningful, so do not spend it unless you are sure. Verdict: try free first, upgrade only if you hit the wall regularly.
If you’re a casual user
You probably do not need Pro. The free tier includes Sonnet, web search, memory, files and Claude Docs. If you only see the limit message once in a while and do not code, save your money. To see where the free ceiling sits before you decide, read Claude free plan limits. Verdict: stay on free.
If you do upgrade and you’re self-employed or run a business, Claude Pro is a tax-deductible business expense. Software subscriptions used for work can qualify as business deductions, but eligibility depends on your situation and this is not tax advice.
Get the Claude Pricing and Plans Cheat Sheet (PDF)
Every Claude plan and the current API rates on one page, plus the gotchas that change what you actually pay. The PDF’s figures were last verified September 27, 2026 and predate this page’s October 26, 2026 re-read, so where the two differ, trust this page.
Which plan should you pick?
Bottom line: most individuals should start on Free and move to Pro at $20/month only after they repeatedly hit the usage limit or need Claude Code, Research or Opus. Max ($100 or $200/month) is for people who outgrow Pro, Team (from $20/seat/month, minimum 2 seats) is for shared work, and Enterprise ($20/seat/month plus usage, 20-seat minimum) is for security and compliance. Answer in order and stop at the first yes.
Are you buying for two or more people, or do you need shared projects, SSO or central billing?
YESTake Team, Standard seats at $20 annual or $25 monthly. Need SCIM, audit logs, HIPAA-ready controls or 20 or more seats? Take Enterprise. 2 to 150 seatsNOGo to question 2.Are you already on Pro and still hitting the five-hour limit more than three times a week, a month in?
YESTake Max. Max 5x at $100 costs the same per unit as Pro; Max 20x at $200 is half that. Pick 20x only if you will use it. $100 or $200NOGo to question 3.Do you need Claude Code, Research, Opus, Claude in Chrome or Microsoft 365, or do you hit the free limit more than a couple of times a week?
YESTake Pro. Every one of those is a dash on Free and a Yes on Pro. $20/mo or $17 annualNOGo to the end.Stay on Free.
Three no answers mean you sit inside what Free gives you and the 1x allowance fits how you work. Start here regardless: it costs nothing to find out, and Anthropic moves a cancelled Pro or Max account back to Free, with your chats, projects and files intact. Compare it feature by feature in Is Claude free?
A verdict for each kind of reader
A student or casual user
FreeSonnet, web search, memory, files and Claude Docs cover everyday work. Anthropic offers no individual student discount.
A writer or analyst using it daily
Free, then Pro at the wallUsage is the gate that bites. Pro is at least 5x Free for $20, and annual billing brings it to $16.67 per unit.
A developer who wants Claude Code
ProClaude Code is in every paid plan and shares the plan's usage pool; heavy users can add pay-as-you-go API credits through a Console account.
A researcher or Microsoft 365 user
ProResearch, Claude in Chrome and Claude for Microsoft 365 and Outlook are No on Free and Yes on Pro.
Someone who lives in Claude all day
Pro, then MaxMax 5x is the same price per unit as Pro, so move up for fewer interruptions; Max 20x is the only plan that halves the price per unit.
A small team
Team StandardPer Pro-sized session, a Standard seat on annual billing is $16.00 against Pro's $16.67, plus sharing and SSO; the floor is two members.
A company with compliance needs
EnterpriseSCIM, audit logs, custom retention and a HIPAA-ready offering start here, at $20 a seat plus usage at API rates and a 20-seat minimum.
A freelancer or sole proprietor
Free, then Pro as a deductionIf Pro pays for itself, keep the receipt: software used for work can count as a business expense, but whether it does depends on your situation. This is not tax advice; check with a tax professional.
A receipt scanner and a tax guide for the self-employed
As an Amazon Associate I earn from qualifying purchases.
Related guides
- Is Claude free?, everything the free tier includes and its limits
- Claude Pro vs Free, detailed feature comparison
- Claude Code review, the coding agent that comes with Pro
- ChatGPT vs Claude vs Gemini, three-way comparison
How this page was read, and how the finding is computed
What was done. On October 26, 2026 every Claude plan, price and usage multiple on this page was read from Anthropic's own pages. We did not run an account on each plan for this revision, so nothing here is a hands-on test and the page carries no “tested” date. Sources, all read that day:
- claude.com/pricing: plan cards, the individual and Team and Enterprise comparison tables, the usage-limit, Claude Code and annual-billing answers, and API prices per model.
- What is the Pro plan?: Pro price, the weekly limit, the priority access wording, the discounts and promotions statement, the API exclusion.
- What is the Max plan?: Max 5x $100, Max 20x $200, monthly billing only, monthly API credits.
- What is the Team plan?: Standard and Premium seat prices, 2 to 150 seats, and the 1.25x and 6.25x Pro per-session multiples.
- Claude Fable models on your plan: Fable on usage credits for Pro and Team Standard seats, and within 50% of weekly limits for Max and Team Premium seats.
- API pricing and Models overview: per-token prices, current and legacy models, tokenizer words-per-token note.
- Google AI plans (US) and ChatGPT pricing: Google AI Plus $4.99, Pro $19.99, Ultra from $99.99; ChatGPT Free $0, Go $8, Plus $20, Pro from $100, all read in a browser October 26.
The finding, step by step. Define one Pro-sized session as Pro's per-session usage allowance. Anthropic's pages give each plan a size: Max 5x is 5 and Max 20x is 20 (“Max gives you 5x or 20x more usage per 5-hour session than Pro”); Team Standard is 1.25 and Team Premium is 6.25 (the Team help article); the pricing page cross-checks the pair, because “Premium seats give 5x more than Standard” and 1.25 x 5 = 6.25. Price per Pro-sized session is the monthly price divided by the size:
- Pro: $20 / 1 = $20.00. Annual: $200 / 12 = $16.67 (Anthropic rounds this to $17).
- Team Standard: $25 / 1.25 = $20.00. Annual: $20 / 1.25 = $16.00.
- Max 5x: $100 / 5 = $20.00. Max is sold monthly only.
- Team Premium: $125 / 6.25 = $20.00. Annual: $100 / 6.25 = $16.00.
- Max 20x: $200 / 20 = $10.00.
- Free: $0, at most 1/5 of a session, because Pro is “at least 5x” Free. Enterprise has no multiple (usage is billed at API rates) and is excluded.
Four of the five plans that publish a multiple land on exactly $20.00. The arithmetic is re-run by a small script that also asserts the cross-check and the Pro annual saving ($240 - $200 = $40); it passed all checks.
What this does not measure. A session multiple is not a total multiple. Paid plans carry weekly limits that Anthropic does not publish as numbers, and Team's weekly limit is not stated relative to Pro. Anthropic may also limit usage “in other ways, such as weekly and monthly caps or model and feature usage, at our discretion.” Treat the ladder as the price of the one number Anthropic prints, not a promise of equal output.
What we could not confirm. ChatGPT and Google plan prices were read in a browser; the competitor cells those pages do not cover (Codex, ChatGPT context window, image and annual rows, Google context window and web search) remain earlier September figures in italics. Which Sonnet and Haiku version the Free plan serves (the table names none). Whether Free has a weekly limit. Whether Pro includes priority access (the plan table says No and the Pro help article says yes). The price of an Education plan, which Anthropic does not list. Whether the help articles' text matches the pricing page on every future date: both change at Anthropic's discretion. Reuse: cite the per-session figures with a link to this page.
Standards and research behind this page
Claims about what an AI tool can and cannot do are easiest to check against the standards bodies and research groups that measure these systems rather than against any vendor's own description. The sources below are the primary ones, linked so a reader can verify a claim here without taking this page's word for it.
- NIST AI Risk Management Framework is the US federal reference for evaluating and governing AI systems, including how their capabilities and limits should be characterised.
- NIST AI RMF Knowledge Base sets out the measurable characteristics of a trustworthy AI system, which is the vocabulary these comparisons use.
- Stanford HAI AI Index publishes the annual measured benchmarks on model capability, cost and adoption that vendor claims are checked against.
- National Science Foundation AI programme funds and documents the underlying research these products are built on.
- arXiv cs.AI carries the primary papers behind most model capability claims, usually months before a product mentions them.
- FTC Endorsement Guides governs how a review or recommendation must disclose a material connection, which is why the disclosure on this page exists.
- US Copyright Office states what copyright covers, which decides who owns the output of a generative tool.
- Section 508 defines the federal accessibility requirements a tool must meet to be usable in public-sector work.
This page compares tools and does not endorse one. Capabilities and pricing change frequently, and a reader should confirm current behaviour with the vendor before relying on it.