Skip to content
Skuto
Free

Skuto Plan Picker

Find out in one minute whether a paid AI plan is worth it for you: real prices, each with its source.

Question 1 of 3

What do you mostly do with AI?

See the full table, all plans compared
All consumer AI plans compared: monthly price, training default and key usage limit
PlanPrice per monthTraining defaultKey limit
ChatGPT (free)FreeTrains on your chats unless you switch it offOpenAI now states: 'Free users have unlimited everyday text chats, subject to abuse-prevention safeguards.' The default model is GPT-5.6 Luna. File uploads, image generation, voice and data analysis each keep their own separate limits, and the instant context window is 27K.
ChatGPT Go€7.99 per monthTrains on your chats unless you switch it offOpenAI states that Free and Go users have unlimited everyday text chats, subject to abuse-prevention safeguards. Go runs on GPT-5.6 Luna and does not get GPT-5.6 Sol; the Think option also uses Luna. Uploads, images, voice and data analysis keep separate limits. 54K instant / 256K reasoning context.
ChatGPT Plus€22.99 per monthTrains on your chats unless you switch it offOpenAI no longer publishes a numeric cap for Plus. Its pricing page says only 'Expanded messages and uploads' and 'Limits apply'; the GPT-5.6 article says reasoning limits 'depend on your plan'. Plus adds GPT-5.6 Sol at Medium and High reasoning, legacy models, and a 54K instant / 256K reasoning context.
ChatGPT Pro 100€102.99 per monthTrains on your chats unless you switch it offMore usage than Plus. GPT-6 Pro and GPT-5.6 Sol Pro share one weekly allowance, so switching between them doesn't give you more. OpenAI doesn't publish the number.
ChatGPT Pro 200€229 per monthTrains on your chats unless you switch it offEveryday chat and GPT-5.6 Sol are listed as unlimited, as long as use stays reasonable. The Pro models are capped: GPT-6 Pro has a weekly limit and GPT-5.6 Sol Pro a daily one, but OpenAI no longer publishes the numbers. If you subscribe now, you get a smaller allowance than Pro 200 had before 29 September 2026, and OpenAI doesn't say how much smaller.
Claude (free)FreeAsks you about training at sign-upAnthropic publishes no message count for Free: a session limit resets every five hours, and how much you get varies with demand, with how long and complex your conversations are, and with the features and model you use.
Claude Pro€22 per monthAsks you about training at sign-upAnthropic no longer publishes a message count: Pro gives 'at least 5 times' the Free usage per 5-hour session, plus a separate weekly cap across all models. Long chats and attachments consume the budget faster; extra usage credits can be bought.
Claude Max (5x)€149.99 per monthAsks you about training at sign-up5 times Pro's usage per 5-hour session, plus a weekly limit across all models. Anthropic publishes no message count: how far it goes depends on conversation length, the model and the features you use.
Claude Max (20x)€299.99 per monthAsks you about training at sign-up20 times Pro's usage per 5-hour session, plus a weekly limit across all models. Anthropic publishes no message count: how far it goes depends on conversation length, the model and the features you use.
Gemini (free)FreeTrains on your chats unless you switch it offSince I/O 2026 Google uses compute-based limits instead of fixed prompt counts: the budget refreshes every 5 hours with a weekly cap on top, and complex prompts or Pro-model use consume it faster.
Google AI Plus€4.99 per monthTrains on your chats unless you switch it offAbout 2× the standard (free) compute-based limits, same 5-hour refresh plus weekly cap.
Google AI Pro€21.99 per monthTrains on your chats unless you switch it offAbout 4× the standard compute-based limits, with broader access to the Pro model.
Google AI Ultra (entry)€99.99 per monthTrains on your chats unless you switch it off5× or 20× the Google AI Pro limits (not the free ones), depending on which Ultra subscription you have, with the widest access to Pro models and agent features. The budget still refreshes every 5 hours until you hit the weekly limit.
Google AI Ultra (top)€219.99 per monthTrains on your chats unless you switch it off5× or 20× the Google AI Pro limits (not the free ones), depending on which Ultra subscription you have, with the widest access to Pro models and agent features. The budget still refreshes every 5 hours until you hit the weekly limit.
Vibe (free)FreeTrains on your chats unless you switch it offMistral publishes no fixed number: its pricing page only says the free plan has limited messages and web searches, and that Pro gives you more of both.
Pro€14.99 per monthTrains on your chats unless you switch it offMistral publishes no absolute caps, only multipliers on its pricing table: Pro gets up to 6x the messages, up to 5x the web searches and up to 40x the image generations of Free, plus all-day coding. Its FAQ gives one fixed figure: 150 Flash Answers a day on Pro.
Microsoft Copilot (free)FreeDoesn't train on your chats by defaultMicrosoft publishes no message cap for free Copilot: it says only that the free version is subject to available capacity and limits, and when you hit one you can upgrade or come back later. Signing in adds chat history, image creation, longer conversations and voice.
Microsoft 365 Personal€10 per monthDoesn't train on your chats by defaultMicrosoft no longer publishes a monthly credit count: its limits page rates chat and deep reasoning as 'Extensive', and image creation, Voice and Vision as 'Standard'. The Researcher and Analyst agents are not included, limits vary by feature and system load, and the AI features are for the subscription owner only.
Microsoft 365 Family€13 per monthDoesn't train on your chats by defaultMicrosoft no longer publishes a monthly credit count: its limits page rates chat and deep reasoning as 'Extensive', and image creation, Voice and Vision as 'Standard'. The Researcher and Analyst agents are not included, limits vary by feature and system load, and the AI features are for the subscription owner only.
Microsoft 365 Premium€22 per monthDoesn't train on your chats by defaultMicrosoft publishes no numbers for Premium any more: its limits page rates chat, image creation and the Researcher and Analyst agents as 'Extensive', and Voice and Vision as 'More' than on Personal and Family. Limits vary by feature and system load, and the AI features are for the subscription owner only.
Perplexity (free)FreeTrains on your chats unless you switch it offUnlimited standard searches; officially 3 Pro Searches per day and 1 Research query per month (September 2026 figures; Perplexity has changed these quotas before, so treat them as movable).
Perplexity Pro$20 per monthTrains on your chats unless you switch it offPro Searches have a weekly limit and Research queries a monthly one, both set for 'average use'; Perplexity publishes no numbers for Pro. Pro has no monthly credits for Computer, Perplexity's agent for multi-step tasks: new subscribers may get a one-time bonus of 4,000 credits that expires after 30 days, and you can buy more.
Perplexity Max$200 per monthTrains on your chats unless you switch it offThe highest limits of the personal plans: weekly Pro Search and monthly Research limits set for 'advanced use', 10,000 Computer credits a month, higher limits for 'Create files and apps' (reports, spreadsheets, dashboards) and the highest limit for the Comet browser agent. Perplexity publishes no search numbers for Max.

Last verified:

3-4 quick questions, about a minute
€ real prices, with a VAT note on every plan
Weekly vendor pages re-read by an automated scan

How to use it

  1. Answer 3-4 questions

    What you use AI for, how often, and what matters most. If it sounds like work use, it also asks who pays. Tap, next, go back anytime.

  2. Read your pick

    One plan, its monthly price, what you get, and whether it trains on your chats by default.

  3. Check the table

    Every plan side by side with price, training default and usage caps. Then open the vendor's own page to confirm.

What it does: an honest "should you pay?"

Tell it what you do with AI, how often, who pays and what matters most. It recommends one plan from the current ChatGPT, Claude, Gemini, Microsoft 365, Vibe (Le Chat) and Perplexity line-ups, with a runner-up, the real monthly euro price, and a plain note on what that plan does with your chats.

Free plans compete on equal footing here. Light use gets a free pick and a straight "you don't need to pay yet". That's the right answer for a lot of people. A tool that always lands on a paid plan is a sales page, not a tool.

Where every price comes from
  • Straight from official sources: each vendor's pricing page (OpenAI, Anthropic, Google, Microsoft, Mistral, Perplexity) or, for some plans, the official app-store price.
  • Euros where the source lists euros, with VAT included only where that price shows it. Where only dollars exist (Perplexity), we show dollars before tax and say so. We never invent a euro figure.
  • An automated scan re-reads those pages weekly. "Last verified" below moves only when we actually checked.

Full process on our how we test page.

No vendor can buy the top spot
  • A small, fixed scoring rule: capacity fit, then your priority (price, privacy, power or best-free), then training default.
  • Vendor links are marked sponsored in case any become affiliate links later; the scoring rule never sees the links, so the ranking can't shift.
  • Usage caps change: vendors tighten them under load. Treat the limits column as a fair guide, not a contract.

Frequent questions

Is the free plan enough for most people? +

Often, yes, and the tool will tell you so. A few questions a week, and every major free tier covers you; the picker's first job is to say "you don't need to pay yet" when that's true. Paying makes sense once you hit caps mid-task: daily work, long documents, coding, lots of images. The questions find that line.

Why are some prices in dollars? +

Perplexity publishes its consumer prices only in US dollars and adds your local tax at checkout, so there is no fixed euro list price. For its plans the picker shows the dollar price before tax and says so, instead of inventing a euro number. Every other plan is shown in euros. Claude sits in between: Anthropic's own pricing page lists dollars, so the euro price we show is the Claude app's price on Apple's App Store, the same in France, Italy, Germany and Spain, VAT included. A short note under each result tells you where the price comes from, when we checked it and, where the source says so, whether VAT is included.

What's the AVV thing for German businesses? +

AVV is short for Auftragsverarbeitungsvertrag, the data processing agreement (DPA) that GDPR requires when a company lets a provider process personal data on its behalf. Consumer AI subscriptions, even expensive ones, don't include one. If a business pays and personal data is involved, the vendor's business product or API is the route that comes with a DPA. The picker flags this whenever you answer "my business pays".

Do paid plans stop training on my data? +

Mostly no, and it's a common misunderstanding. Big consumer paid plans keep the same default as their free tiers: on unless you switch it off in settings. There are exceptions: Anthropic asks you to choose at signup. Mistral is not one of them: its paid Vibe plans (Pro and Team) train by default too, and only its Enterprise plan starts opted out. The picker shows the actual default for every plan it recommends.

How fresh are these prices? +

Every price comes from an official source: the vendor's own pricing page or, for some plans, its official app-store price. VAT counts as included only where that source shows it, and an automated scan re-reads the vendor pages weekly. The "Last verified" date at the bottom of this page moves only when we actually re-checked. We never bump dates to look fresh. AI pricing changed several times this year alone, which is exactly why the dates are part of the tool.

Set location

Tunes each tool to where you are. Stays on this device.

Set your AI

Which AI do you use?

Not sure? Help me choose →