ChatGPT Plus
OpenAI
$20/month
See limits and details
ChatGPT, Claude, Gemini, Cursor and more. Answer three quick questions and see the plans that fit you, with the price and the usage limit of each.
Question 1 of 3
Three questions. Nothing to type.
An individual plan from each of the 8 services, with its monthly price and where we read it. Open a card for its usage limit and what it means in practice.
OpenAI
$20/month
See limits and details
Anthropic
$20/month
See limits and details
$19.99/month
R$ 96.99/month · charged in Brazil
See limits and details
Perplexity AI
$20/month
See limits and details
Anysphere
$20/month
See limits and details
GitHub
$10/month
See limits and details
xAI
$30/month
See limits and details
Mistral AI
$14.99/month
See limits and details
Tick the subscriptions you have today. You get the monthly and yearly total, and a note where two plans cover the same job.
Open the bundle calculatorOne row per plan. Next to each price and limit you see where we read it and on which day; a figure from a third party says so.
Scroll sideways to see every column.
| Plan | Price per month | What the source says about limits | What it means in practice |
|---|---|---|---|
| ChatGPT GoOpenAI | $8/month third-party · checked · cloudzero.com (updated Sep 4, 2026) | Not published in a source we can cite | The cheapest paid ChatGPT tier. We could not read its limits on an official page, so treat it as "more than Free, less than Plus" and check the pricing page before paying. |
| ChatGPT PlusOpenAI | $20/month third-party · checked · cloudzero.com (updated Sep 4, 2026) | Roughly 160 messages per three hours on the main model, plus Deep Research tasks and Agent mode. third-party · checked · cloudzero.com (updated Sep 4, 2026) | Enough for a working day of normal chat for most people. The number comes from a third-party article, and OpenAI has changed it before without notice. |
| ChatGPT ProOpenAI | from$100/month third-party · checked · cloudzero.com (updated Sep 4, 2026) | $100 tier: 5x Plus usage. $200 tier: 20x Plus usage, about 250 Deep Research tasks a month and priority compute. third-party · checked · cloudzero.com (updated Sep 4, 2026) | Only worth it if you hit the Plus cap most days. If you hit it once a week, waiting is cheaper than paying five times more. |
| ChatGPT BusinessOpenAI | $25/month · per seat third-party · checked · cloudzero.com (updated Sep 4, 2026)$20/month billed annually | Not published in a source we can cite | A team plan billed per seat. The per-person price is close to Plus; what you pay for is shared workspace and admin control, not more messages. |
| Claude ProAnthropic | $20/month official · checked · claude.com/pricing$17/month billed annually | Usage limits reset on a rolling five-hour session window; paid plans add weekly limits on top. official · checked · claude.com/pricing | There is no fixed message count. A long session can use up the five-hour window, then you wait for it to reset; the weekly limit caps heavy use across the week. Claude Code draws from the same allowance. |
| Claude MaxAnthropic | from$100/month official · checked · claude.com/pricing | Choose 5x or 20x more usage than Pro. Same five-hour window and weekly limits, with higher output limits and priority access at high-traffic times. official · checked · claude.com/pricing | The page shows "From $100" for the 5x option and did not show the 20x price in the text we read. Meant for people who run Claude Code for hours a day; a casual user will not notice the difference from Pro. |
| Claude TeamAnthropic | $25/month · per seat official · checked · claude.com/pricing$20/month billed annually | Standard seat: $25 per month, or $20 billed annually. Premium seat: $125 per month, or $100 billed annually. Seat types can be mixed. official · checked · claude.com/pricing | You can give premium seats only to the people who need heavy usage and keep everyone else on standard seats, which is usually cheaper than putting the whole team on Max. |
| Google AI PlusGoogle | No USD price we can cite R$ 24.99/month · charged in Brazil official · checked · gemini.google/subscriptions (Brazil) | Usage limit 2x higher than the free version; 200 monthly Flow credits; 400 GB of storage. official · checked · gemini.google/subscriptions (Brazil) | A small step up from free Gemini. Third-party sites disagree on its US price ($4.99 or $7.99), so we do not list one. |
| Google AI ProGoogle | $19.99/month third-party · checked · aipricing.guruR$ 96.99/month · charged in Brazil official · checked · gemini.google/subscriptions (Brazil) | Usage limit 4x higher than the free version; 1,000 monthly Flow credits; 5 TB of storage; priority access to the Pro model. official · checked · gemini.google/subscriptions (Brazil) | "4x the free limit" is relative to a number Google does not publish. The concrete parts are the 5 TB of storage and the Flow credits: if you already pay for Google storage, part of the price is covered. |
| Google AI UltraGoogle | No USD price we can cite fromR$ 779.90/month · charged in Brazil official · checked · gemini.google/subscriptions (Brazil) | From R$ 779.90 (5x limits) or R$ 999.90 (20x limits) in Brazil; usage limits up to 20x higher than Pro; 10,000 to 25,000 Flow credits; from 20 TB of storage. official · checked · gemini.google/subscriptions (Brazil) | Priced for people who generate video daily. Third-party sites disagree on the US price, so we show only what Google charges in Brazil. |
| Perplexity ProPerplexity AI | $20/month third-party · checked · cloudzero.com (Sep 21, 2026) | Not published in a source we can cite | A search-first assistant: answers come with sources. We found no official number for its daily search cap, so we do not print one. |
| Perplexity MaxPerplexity AI | $200/month third-party · checked · cloudzero.com (Sep 21, 2026) | Not published in a source we can cite | Ten times the price of Pro. The extra is agent features and credits, not better search answers, so most researchers stay on Pro. |
| Cursor ProAnysphere | $20/month official · checked · cursor.com/pricing | Extended limits on Agent. After the included amount is consumed, on-demand usage is billed in arrears. official · checked · cursor.com/pricing | The included usage is a budget, not a message count: expensive models spend it faster. With on-demand usage switched on, a heavy month costs more than $20. |
| Cursor Pro+Anysphere | $60/month official · checked · cursor.com/docs | More included usage than Pro. The model you pick changes how quickly included usage is consumed. official · checked · cursor.com/docs | The middle step for people whose Pro allowance runs out mid-month. Compare it with what you actually paid in on-demand charges on Pro. |
| Cursor UltraAnysphere | $200/month official · checked · cursor.com/docs | The largest included usage of the individual plans. The docs do not state a dollar amount. official · checked · cursor.com/docs | For developers who keep agents running most of the day. If your on-demand bill on Pro stays under $180 a month, Ultra costs more than what you already pay. |
| Cursor TeamsAnysphere | $40/month · per seat official · checked · cursor.com/pricing | $40 per user per month. A Premium seat at $120 per user gives 5x the Standard limits on Agent. official · checked · cursor.com/docs | Twice the price of Pro per person. The difference buys central billing, team-wide privacy mode and SSO. |
| Copilot ProGitHub | $10/month official · checked · docs.github.com | 1,500 AI credits a month (1,000 base plus 500 flex). Unlimited code completions. official · checked · docs.github.com | Autocomplete never runs out; chat and agent requests spend credits, and stronger models spend more per request. The cheapest paid coding plan in this guide. |
| Copilot Pro+GitHub | $39/month official · checked · docs.github.com | 7,000 AI credits a month (3,900 base plus 3,100 flex). official · checked · docs.github.com | About 4.7 times the credits of Pro for 3.9 times the price. It pays off only if you use up the Pro credits every month. |
| Copilot MaxGitHub | $100/month official · checked · docs.github.com | 20,000 AI credits a month (10,000 base plus 10,000 flex). official · checked · docs.github.com | GitHub describes it as a plan for sustained, high-volume use. At this price it competes with Claude Max and Cursor Ultra rather than with Copilot Pro. |
| Copilot BusinessGitHub | $19/month · per seat official · checked · docs.github.com | 1,900 AI credits per user per month. Enterprise: $39 per seat with 3,900 credits. official · checked · docs.github.com | Slightly more credits than Pro per person, with organisation-level policy control. Billed per granted seat, so unused seats still cost money. |
| SuperGrokxAI | $30/month third-party · checked · cloudzero.com (updated Sep 1, 2026) | Paid plans share one weekly usage pool spent across Chat, Imagine, Voice and Build. xAI publishes no fixed caps. third-party · checked · cloudzero.com (updated Sep 1, 2026) | One weekly pool means a heavy image session leaves less for chat that week. Other sites list extra tiers (Lite, Plus) that our source does not mention, so we leave them out. |
| SuperGrok HeavyxAI | $300/month third-party · checked · cloudzero.com (updated Sep 1, 2026) | Same weekly pool as SuperGrok, with a larger allowance. No published figure. third-party · checked · cloudzero.com (updated Sep 1, 2026) | The most expensive individual plan in this guide, with no published limit to compare against. Hard to justify without trying SuperGrok first. |
| Le Chat ProMistral AI | $14.99/month official · checked · mistral.ai/pricing | More messages and web searches than Free; all-day coding in the CLI, IDE or on the web; $15 a month in API credits. Price excludes taxes. official · checked · mistral.ai/pricing | The lowest-priced full assistant here, and the $15 of API credits nearly equals the subscription if you also build with the API. "More messages" has no published number. |
| Le Chat TeamMistral AI | $24.99/month · per seat official · checked · mistral.ai/pricing | $24.99 per user per month, excluding taxes. Up to 30 GB of storage per user. official · checked · mistral.ai/pricing | Priced like the other team plans. The differences listed are storage, domain verification and data export. |
Every price and limit here shows its source and the date we checked it. No affiliate links: the order comes from the table, not from who pays.
Six things you can check for yourself on this site.
65 figures across 24 plans from 8 services, each stored with the page it was read on and the day it was read.
48 figures were read on the provider’s own page. The other 17 come from named third-party articles, because those providers block automated reading, and they are marked as unconfirmed.
A figure not re-checked for more than 30 days is flagged on the page, and the project’s test suite fails until someone re-reads the source.
The recommendation is a filter over 24 written rows, printed in full on the quiz page. No score, no weights.
Outbound links go to the providers’ own pricing pages, with no referral code. The order of results comes from the table.
Where a provider charges in Brazilian reais we show that price. Everything else is a conversion with the rate, its date and the tax shown.
The price of an AI plan is easy to compare. The limit is not, because providers describe it in different units. Reading the unit first tells you how the plan will behave on a busy day.
“About 160 messages every three hours.” You can plan around it: when the count is spent, you wait for the period to end. Long and short messages cost the same.
“Limits reset on a five-hour window, with a weekly limit on top.” There is no message count. Long conversations and large files use the window faster, so one deep session can end your afternoon while light use never notices the limit.
“1,500 credits a month” or “included usage”. Each request costs a different amount depending on the model. A stronger model can spend in a week what a lighter one spends in a month, and some plans keep going after the budget ends and bill the difference.
“5x more than Pro” or “4x the free limit”. This only tells you the plan is bigger. Unless you already know how often you hit the smaller plan’s limit, the multiple cannot be turned into a number, and we do not invent one.
Two plans at a time: price, limits, what is included and where each figure comes from.
It depends on what you do with it. ChatGPT Plus, Claude Pro, Perplexity Pro and Cursor Pro all cost $20 a month, but they limit usage in different units: messages per three hours, a five-hour window with a weekly cap, or a usage budget that expensive models spend faster. Answer the three questions at the top of this page and read the reason given for each plan.
From the pricing page of each provider, read on the date shown next to each figure. ChatGPT, Perplexity, Grok refuse automated reading of their pricing pages, so their figures come from a named third-party article and are labelled as not confirmed on the official page.
Every figure carries the date it was last read. A figure older than 30 days is flagged on the page, and the project's own test suite fails until someone re-reads the source. A scheduled job also hashes the pricing pages that allow it and raises an alarm when the text changes.
Because the provider does not publish one. Several plans describe their limit only as a multiple ("5x more than Pro") or as "more usage". We print the wording of the source and then say, in a separate column, what it means in practice. We do not fill the gap with numbers from forums.
No. There are no affiliate links on this site: links to providers go to their pricing pages without any referral code. The quiz order comes from a written decision table that is published in full.
No. This guide covers paid subscriptions. Free tiers are a separate comparison.