Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
Latest news
Editions
The best AI chatbots in November 2026 differ less on raw intelligence than on how long you can keep using them before a meter cuts you off. Notebookcheck’s early November subscription roundup found that plans around $20 differ mainly in their extras, while top tiers start at about $100. That gap is where most readers burn money, because two plans with the same price can run out in an afternoon or last a full working week.
This comparison ranks ChatGPT, Claude, Gemini, Grok, Perplexity and Microsoft Copilot on one question: which paid plan runs out first, and how does each vendor tell you? Prices below are list prices published in US dollars. Irish buyers should expect euro pricing and VAT at checkout, so treat every figure as a guide and confirm it on the vendor’s own page before you subscribe. Data is current as of 7 November 2026.
One boundary first. Attachment size and page-count ceilings are a separate subject, and our file upload limits comparison covers them. Here we look at message allowances, time windows, weekly pools and credit systems, which is what most people mean when they say a chatbot ran out.
The short version of the findings: no vendor publishes a clean, comparable message count for its consumer plans. Anthropic uses five-hour windows. xAI sells multiples of its cheapest paid plan. Google sells a 5x multiplier between its Pro and Ultra tiers. OpenAI splits access by model and by feature. Microsoft publishes no comparable numeric quota for Ireland at all. Picking among the best AI chatbots therefore means understanding the metering model, not just reading a price tag.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
Every major chatbot now meters usage, but they meter different things. Some count messages. Some count compute. Some count wall-clock time since your last prompt. A few fold coding agents, image generation and voice into one pool. Knowing which of these your favourite assistant uses matters more than comparing headline model names, because the same person can sit well under the cap on one service and get locked out of another on the same day.
Anthropic is the clearest example. Claude’s subscription limits work in five-hour windows, and on 31 August 2026 the company raised those limits for Pro, Max, Team and seat-based Enterprise plans. It also added a rate-limit reset that users can save and trigger themselves. A window system rewards bursts. You can work hard for a stretch, hit the wall, and come back once the window rolls over. It punishes people who need steady all-day output, because the wall arrives at an unpredictable point in the afternoon.
OpenAI’s business documentation uses a similar idea. The official Business limits page gives model-specific estimates per five-hour window rather than one universal monthly cap. That wording matters. It means the answer to how many messages you get depends on which model you pick, and the heavier reasoning models drain the allowance faster than the light ones.
xAI took a different route. Its documentation says paid Grok plans raise limits across products through a shared weekly allowance, and the plan ladder is expressed as multiples of the cheapest paid tier. Premium lists 4x the usage of Plus and Super lists 15x. There is no per-day message count, which is easier to market and harder to plan around. Google does something similar between its Pro and Ultra consumer plans, describing Ultra as providing 5x higher usage limits in Gemini and Antigravity than AI Pro.
The third model is the meter you can top up. Grok users can buy extra usage credits starting at $5. Perplexity’s top consumer plan, Max, is reported to include 10,000 Computer credits per month. Microsoft said on 25 September 2026 that long-running agent features such as Cowork, Code and Autopilot, along with frontier models such as Astra and Fable, would use usage-based billing rather than being covered by a standard licence. Each approach shifts the risk differently, and none is as predictable as the old flat-rate subscription.
A limit is also not an outage. If a service is simply down, no plan tier helps, and our outage reliability comparison tracks that separately. What follows is about the planned ceiling you hit when everything works.
The table below lists each plan with its limit mechanism and the only numbers confirmed on vendor pages or by at least two independent sources. Where a vendor does not publish a figure, the cell says so rather than guessing. That blank is itself a finding. It tells you how much trust to place in any third-party claim about exactly how many messages a plan allows.
Two patterns jump out. First, only Anthropic and OpenAI describe time-boxed windows at all, and neither prints a message total for consumer plans. Second, the multiplier approach used by xAI and Google tells you how plans compare to each other but not how much you get in absolute terms. A 15x multiplier on an unknown base is still an unknown. Read the last column as a map of what each vendor is willing to commit to in public, and you will see why independent testing matters.
OpenAI runs the most layered ladder of the six. According to ChatGPT’s own pricing page, the consumer plans are Free, Go, Plus and Pro, with Business and Enterprise for organisations. Free and Go users get standard access on GPT-5.6 Luna, while higher plans receive larger allowances plus additional reasoning and Pro models. Free users get unlimited everyday text chats, subject to abuse-prevention safeguards, but file uploads, image generation, voice and data analysis each carry their own limits.
The official model matrix shows where the walls sit. GPT-6.1 Sol is marked unavailable on Free and Go, available on Plus, and expanded on Pro. GPT-5.6 Sol follows the same pattern, with Pro showing Unlimited with an asterisk. GPT-5.6 Sol Pro, the higher-capability version for demanding work, is listed only for Pro. In practice a Plus subscriber does not run out of chat. They run out of the best model and the heaviest features first.
The Pro line is where OpenAI has been most active. The Pro 500 tier includes Ultrafast access to GPT-6 Astra and GPT-6.1 Sol, and OpenAI states it has the highest included usage of the three Pro tiers. GPT-6 Pro, powered by GPT-6 Astra, is available on Pro 100, Pro 200, Business and Enterprise. A single-source listing from 29 October put those tiers at $100, $200 and $500 per month, so treat the exact prices as reported rather than confirmed. The same reporting says Pro 200 users saw their allowance change after 29 October, while another source says an announced allowance cut was cancelled and Pro 200 reopened to new subscribers. The sources conflict, and that conflict is a warning in itself: check the allowance on the day you subscribe.
Business is the plan with the most explicit numbers, though the figures come from a single source. Business Standard reportedly includes 15 GPT-6 Pro messages per month, while Business Premium reportedly includes 50 Pro messages per week, shared across GPT-6 Pro and GPT-5.6 Sol Pro. OpenAI’s help centre confirms Premium seats cost $125 monthly or $100 on annual billing. For a team that leans on the Pro models, a 15-per-month cap is tight enough to matter inside the first week.
Model churn also affects limits. OpenAI is reportedly retiring GPT-5.5 from ChatGPT, ChatGPT Work and Codex, and the GPT-6 rollout began reaching eligible paid plans around 6 November. When models change, allowances change with them. If you are weighing the lower paid tiers against rivals on price alone, our ChatGPT Plus, Claude Pro and Perplexity Pro price comparison sets out the monthly cost gap, and our Gemini 4 Argon vs Claude Opus 5.5 vs GPT-6.1 Sol analysis covers how the flagship models themselves compare.
Anthropic’s current lineup is Claude Opus 5.5, Claude Sonnet 5.5 and Claude Haiku 5.5. Anthropic says the newest Sonnet is more than 30% faster than Sonnet 5 and can cost up to 30% less for most work, and that its coding-focused model runs 40% cheaper than Opus 5 on typical workloads. Those cost cuts matter for limits because a cheaper model to serve is a model Anthropic can afford to give you more of. Haiku 5.5 reportedly arrived on 7 October, with Anthropic claiming it costs about 75% less to run than Haiku 4.5.
On Claude’s pricing page, Pro costs $20 per month billed monthly, or $17 per month when billed annually ($200 up front). Max starts from $100 per month. Team seats are $25 monthly or $20 annually, and the page lists weekly limits on premium seats, with one line describing 50% of weekly limits on those seats. So Claude is not purely a five-hour system. Window limits sit on top, and weekly caps apply to some seat types.
The 31 August change is the single most important limits event of the quarter. Anthropic raised five-hour usage limits across Pro, Max, Team and seat-based Enterprise, and added the user-saveable reset. For anyone who left Claude over limits earlier in 2026, this is the change that should prompt a second look. The saved reset also gives heavy users a lever the others lack: you decide when to spend it instead of waiting for the window to roll.
Developers should note the API side separately. The Opus 5.5 API price is $4 per million input tokens and $20 per million output tokens, with a fast mode at $8 and $40. Anthropic’s pricing page also lists up to 2.5x faster speeds with fast mode for Opus 5.5 at twice the standard price, and 50% savings with batch processing. Those are pay-per-token rates with no five-hour window, which makes the API the escape hatch when a subscription cap blocks urgent work. Claude Mythos 5.1, priced from $10 and $50 per million tokens, is restricted to vetted cyberdefenders and life-science researchers, so an ordinary subscriber cannot reach it.
One more item for Irish and EU readers. Microsoft announced on 5 to 6 November that Claude models are coming into the EU Data Boundary inside Microsoft Copilot. That gives European organisations a second route to Claude, through Microsoft’s licensing and usage rules instead of Anthropic’s own subscription. A single-source report adds that Anthropic’s updated 2026 usage policy is scheduled to take effect on 12 November, so it is not yet in force as of this writing.
Google sells Gemini through Google AI Pro and Google AI Ultra. Its newest consumer-facing model is Gemini 4 Argon, which became globally available on 13 October after a controlled release to trusted cyber defenders on 30 September, per single-source reporting. Gemini 4 Carbon, focused on long-horizon engineering and cybersecurity, was announced on 28 October but is not confirmed as generally available. Gemini 3.8 Flash is the newest Flash model and reaches AI Pro and Ultra subscribers in the Gemini app, AI Mode in Search and Gemini in Sheets.
The limit story is the multiplier. Ultra was announced at $100 per month and described as giving 5x higher usage limits in Gemini and Antigravity than AI Pro. Google has not published a per-day prompt count for Pro, so the 5x figure has no absolute anchor. Reports suggest Argon in the Gemini app reaches Ultra users and only some Pro users, which means a Pro subscriber may find the flagship either missing or rationed. Check what your own account shows before you assume you have it.
Gemini’s strength is surface area. Google lists access through the Gemini app, AI Mode in Search, Gemini in Sheets, Google AI Studio, the Gemini API, Antigravity, Vertex AI, Gemini Enterprise and NotebookLM. If your work already lives in Google’s tools, usage may feel less metered because some of that activity happens inside products you already pay for. Our app integrations comparison maps those connections in detail. On 5 November Google reportedly said Gemini can now generate, preview and edit documents, PDFs, presentations and spreadsheets in the app, which will pull more document work into the same allowance.
For developers, Argon’s API has a promotional price of $2 per million input tokens and $10 per million output tokens through 30 November 2026. From 1 December the stated price becomes $4 and $20. Anyone planning to build on Argon should budget at the higher rate. For schools, a reported 3 November Google update said access for K-12 student accounts stays disabled unless administrators switch it on, which is worth knowing for Irish second-level settings.
xAI restructured Grok on 12 October 2026 into Free, Plus, Premium, Super and Ultra. Free costs nothing and gives daily Grok access with limited usage. Plus is $8 per month for more Grok and Cursor usage. Premium is $30 per month and includes the benefits of SuperGrok, Cursor Pro and X Premium+, which xAI says replaces a combined cost of $90 per month. Super is $100 per month at 15x the Plus allowance, and a single-source listing puts Ultra at $200 per month and 50x.
This is the most transparent ladder in relative terms and the least transparent in absolute ones. xAI states the multiples (4x, 15x, and reportedly 50x) but gives no message quota, reset period or model-specific cap. The allowance is measured through one shared usage pool, and extra credits start at $5. A single-source report says the pool moved to a shared weekly allowance earlier in the year, replacing separate daily limits for Chat, Imagine, Voice and Build.
The pool design has a practical consequence. If you spend heavily on image generation or coding agents early in the week, you have less left for chat on Friday. People who use only one feature may prefer a per-product limit, while people who rotate between features gain flexibility. Image users should read our image generation price comparison, since Grok’s Imagine usage competes with chat inside that same pool.
On the model side, Grok 4.7 is the current flagship. It has a 500,000-token context window, accepts text and images, outputs text, and supports configurable reasoning effort. Its API price is $2 per million input tokens and $6 per million output tokens for shorter prompts, with a May 2026 knowledge cutoff. xAI also announced Grok Voice 3.0 on 14 October. One reported change lands today: xAI’s documentation says outputs from several models, Grok 4.7 included, are watermarked by default with a per-model opt-out, which matters to anyone using Grok text commercially.
Perplexity is the odd one out because it meters searches rather than chat turns. Its free tier is called Standard, and a single-source description puts it at basic searches, three Pro Searches per day and one Research query per month. Pro costs $20 per month and Max costs $200 per month. Education Pro is reportedly $10 per month after SheerID verification. Pro adds higher Pro Search and Research access, file uploads, a choice of advanced models, image and video generation, and limited weekly Computer usage, according to Perplexity’s own description.
Perplexity’s October 2026 material says the service added a choice of advanced models from OpenAI and Anthropic. A 27 October snapshot lists GPT-6.1 Sol, Gemini 3.8 Flash, Claude Sonnet 5.5, Kimi K3, GLM 5.3 and Grok 4.7 for Pro and Max, with Max also listing GPT-6 Astra, Gemini 4 Argon, Claude Opus 5.5 and Claude Fable 5.1. The company warns that models may be added or retired as providers change availability, so the list is a dated snapshot. For research with sources, the practical result is that Perplexity Pro lets you reach several rival models under one $20 plan, with the search quota as the main constraint. On freshness, our news accuracy comparison tests how each handles recent events.
Copilot is harder to pin down because Microsoft now uses the name across consumer Microsoft 365, commercial licensing, Cowork, Copilot Studio and GitHub Copilot. The redesigned consumer Copilot, announced on 25 September around Home, Code and Autopilot, was scheduled to roll out worldwide in November. Microsoft states that Microsoft 365 Pro includes Copilot with the highest usage for AI features among the consumer tiers, but the reviewed pages give no numeric quota. On the business side, Microsoft confirmed Copilot Business stayed at US$21 per user per month, paid yearly, from 1 July, and its Business Premium plus Copilot Business bundle was US$43. The Microsoft Copilot blog post on agents in business processes is the official source for the usage-based billing model that now sits on top of those licences.
Microsoft also changes models often enough to affect allowances. A single-source report says GPT 6.1 Sol replaced GPT 6 Sol in Copilot Cowork on 7 November and that Grok 4.7 was added for long documents and multi-step work. Another notes Copilot Studio retiring GPT-5 Chat, GPT-5.6 Sol and GPT-5.6 Luna on 12 November, with an EU/EFTA exception for GPT-5 Chat until 30 November that the documentation ties to EU AI Act requirements. If your organisation runs on Copilot Studio, watch those dates.
Price and limits move together, so the table below lines up the tiers where a figure was confirmed. Cells marked Not confirmed mean the vendor page we fetched did not show an Ireland price and no second source backed the number. We chose a blank over a guess.
Reported figures outside the table: Google AI Pro at $19.99 per month and ChatGPT Plus at $20 per month, both from single-source listings. If those hold, the main paid tier at Anthropic, OpenAI, Google and Perplexity converges on roughly $20, which is exactly why limits and extras, not price, separate them. Grok is the outlier with an $8 entry plan and a $30 mid tier that bundles Cursor Pro and X Premium+.
Looking at the second rung, the picture shifts. Claude Max from $100, Gemini AI Ultra at $100 and Grok Super at $100 cluster together, and ChatGPT’s Pro 100 tier is reported at the same price. That $100 rung is the new default for heavy individual users in late 2026. Perplexity jumps straight from $20 to $200, which leaves a gap for people who want more than Pro but not twice the price of its rivals. For a wider look at alternatives to the market leader, see our ChatGPT alternatives guide.
List prices in dollars are a starting point. In Ireland you will usually see a euro price with VAT included, and the exchange rate on the day can move the total by more than the gap between two budget plans. Check the checkout page, not the marketing page. If you invoice a business, confirm whether the plan gives you a VAT invoice, because Business and Team tiers often do and personal tiers may not.
Data location is the second check. Microsoft’s move to bring Claude inside the EU Data Boundary, and its documentation tying one model retirement timetable to EU AI Act requirements, show that European rules already shape which models appear in which plans. A model you see in an American screenshot may arrive in Ireland later, or on a different tier. If a feature matters to your work, test it in your own account during any free trial or on a monthly plan before committing to a year.
Because no vendor publishes comparable message counts, the strongest evidence comes from outside trackers and from reading the vendors’ own pages side by side. Three source types agree on the direction of travel, even where details differ. We found no independent benchmark that counts messages per plan, so treat any article that claims one as the result of its own testing, not an industry standard.
First, Notebookcheck’s price review found that mid-tier plans around $20 differ mostly in extras, and that top tiers start at around $100. It also covers cheaper options such as ChatGPT Go, Google AI Plus and Microsoft 365 Personal, plus what you get without paying. That lines up with the pattern in our own table: the entry and mid tiers look similar, and the real fork is at the $100 rung.
Second, FourWeekMBA’s AI subscription plans tracker frames the change as the end of the flat-rate era. Its sections cover a taller plan ladder and limits that moved to meters, together with a timeline of plan changes. That description fits what the vendors did this year. Anthropic added resets and raised windows, xAI folded daily limits into a weekly pool, Microsoft tied agents to usage-based billing, and OpenAI split Business seats by Pro-message counts.
Third, the vendors’ own pages act as the check on the other two. OpenAI’s pricing page lists per-model access by plan, Anthropic’s lists seat prices and weekly limits on premium Team seats, xAI’s lists multiples of Plus, and Microsoft’s says Pro has the highest consumer usage without giving a number. Hello Growth’s November AI news roundup, which lists 13 developments, tracks plan changes across vendors in the same period. Across all three source types the conclusion holds: plans are getting more tiers, more meters and fewer promises.
The timeline shows how fast the ground moves. Five major limit or pricing changes landed in about ten weeks, and a sixth is already scheduled. Any ranking of limits, including this one, has a short shelf life. That caution applies to our own table too, since several vendor pages gave only partial extracts and some regional prices were not visible to us. The fix is a short personal test, which the migration section describes.
The following are illustrative scenarios, not case studies of named people. Each one uses only the limit structures described above, and none makes up a message count.
1. The thesis writer in Galway. A postgraduate reads dozens of papers and drafts chapters with long pasted excerpts. Claude’s five-hour windows suit this pattern, since work happens in long sessions with natural breaks. The risk is hitting the wall mid-chapter, so the saveable reset is the feature to learn. ChatGPT Plus is a fair alternative, but the best model is the first thing to be rationed.
2. The freelance developer in Cork. Agentic coding eats allowance faster than chat. On Claude, the question is whether Pro’s windows hold through a long session or whether Max from $100 is needed. On Grok, Premium at $30 bundles Cursor Pro, which makes it the cheapest route to a coding agent, but the shared weekly pool means a heavy Monday leaves less for Friday. On Copilot, Code and Autopilot move to usage-based billing, so costs scale with work done.
3. The small agency in Dublin. Five seats, shared client work, regular use of the top reasoning model. ChatGPT Business Standard reportedly caps GPT-6 Pro at 15 messages per month, and Premium at 50 per week. If those figures hold, the team either buys Premium seats at $125 monthly or $100 annually, or accepts that the Pro model is a scarce resource kept for the hardest tasks.
4. The researcher who needs sources. Perplexity Pro at $20 gives search-first answers with a choice of models. The binding limit is Pro Searches and Research queries, not chat turns. Someone running literature scans will want Pro, not Standard. The free tier’s reported three Pro Searches per day and one Research query per month is a taster, not a working allowance.
5. The household user in Limerick. Questions about recipes, travel and homework help. ChatGPT Free’s unlimited everyday text chats under abuse safeguards mean most people never meet a cap. Image generation and voice are where they will, because those carry separate limits. Paying makes sense only if you hit those specific walls.
6. The office worker on Microsoft 365. Copilot Chat reportedly comes with an eligible Microsoft 365 subscription at no extra charge. A person who stays inside Word, Excel and Outlook may never think about limits. The moment they push into long-running agents, usage-based billing applies. An IT admin should ask which side of that line the staff will live on.
These picks follow from the metering models, not from benchmark scores. Where two services are close, we say so.
If you must pick one service and hate surprises, we lean to Claude Pro for its documented window and reset, with ChatGPT Plus as the runner-up when you want the broadest feature set. If your priority is the lowest entry price with a path upward, Grok Plus at $8 is the cheapest paid door in this comparison.
Most users hit a limit because of habits, not volume. Start new chats for new topics, since a long thread forces the model to reprocess its whole history on each turn. Pick the lighter model for routine drafting and save the flagship for the step that needs it. Combine several small questions into one well-formed prompt. Paste only the relevant section of a long document instead of the whole file. These habits cost nothing and often move the wall by a large margin on any of the six services.
When you do hit the wall, you have four options: wait for the reset, use a saved reset where one exists, switch to another assistant for the afternoon, or move the job to a pay-per-token API. Keeping a free account on a second service is the cheapest insurance. Because ChatGPT Free offers unlimited everyday text chats under safeguards, it makes a sensible backup for the days your paid plan runs dry.
People often switch after hitting a limit. A planned switch beats a panicked one. The steps below work between any two of the six services, and the first one doubles as a way to test limits for yourself before you pay.
Because vendors do not publish comparable counts, your own log is the best data you will get. Record how many prompts you send and the time you hit any warning. The small script below appends one line per event to a CSV file you can sort later.
Save your custom instructions as plain text first. Then copy the system prompts and project descriptions you rely on into a notes file. Conversations rarely transfer between services, so paste only the context a new chat needs: a summary of the project, the style rules, and the three or four documents that matter. Do not try to move every old thread.
Recreate your projects, custom instructions and any connectors you use before you cancel the old plan. Check whether the new service reaches the apps you rely on. Run your five most common tasks on both services side by side for a few days and note which one hits a wall first.
Annual billing locks you in, and Claude’s $17 rate is the annual one. If you suspect you may switch again, pay monthly for the first month. Also check upcoming dates before you move: Gemini 4 Argon’s API promotion ends on 30 November, and OpenAI and Microsoft are retiring several models this month. For wider AI coverage, browse our AI and machine learning section.
ChatGPT. Pros: unlimited everyday text chats on Free, the widest plan ladder from Go to Pro 500, and a clear model matrix on the official page. Cons: allowances vary by model and feature, reported allowance changes on Pro 200 have been confusing, and Business Pro-message caps look low.
Claude. Pros: a defined five-hour window, a raised ceiling since 31 August, a saveable reset, and cheaper-to-run models in the 5.5 line. Cons: weekly limits on some Team seats, no printed message total, and heavy use pushes you toward the $100 Max tier.
Gemini. Pros: deep integration across Google products and a clear 5x step to Ultra. Cons: no absolute Pro quota, uncertain Argon access on Pro, and Gemini 4 Carbon still not generally available.
Grok. Pros: the cheapest paid entry at $8, a $30 tier that bundles Cursor Pro, and a 500,000-token context window. Cons: usage published only as multiples, a shared weekly pool that can drain early, and plan names that changed on 12 October.
Perplexity. Pros: search-first answers, a choice of several vendors’ models and a cheap $10 education route (reported). Cons: a $20 to $200 jump with nothing between, search quotas that bite quickly on the free tier, and a model list that changes without notice.
Copilot. Pros: tight Microsoft 365 integration, Copilot Chat reportedly included with eligible subscriptions, and Claude arriving in the EU Data Boundary. Cons: the name covers many products, no numeric consumer quota, and usage-based billing for agents makes costs hard to forecast.
On the evidence available as of 7 November 2026, no chatbot wins on limits outright, because nobody publishes the number you need. What the data does support is a ranking by predictability. Claude comes first because its five-hour window, the 31 August increase and the saveable reset are documented and consistent. ChatGPT comes second because its free tier is generous and its model matrix is explicit, though allowances shift with each model launch. Grok comes third on transparency of the ladder, with the 4x and 15x multiples, but loses points for having no absolute number.
Gemini and Perplexity sit in the middle. Gemini’s 5x Ultra gap is clear but unanchored, and Perplexity’s search quotas are clear but narrow. Copilot ranks last for pure predictability, since Microsoft publishes no numeric quota and long-running agents use usage-based billing. That does not make it a bad product. It means a Copilot buyer needs to model costs more carefully than anyone else.
Our recommendation by budget: at $0, start with ChatGPT Free. At $8, Grok Plus. At about $20, Claude Pro for session work or Perplexity Pro for research. At about $100, choose between Claude Max, Gemini AI Ultra and Grok Super based on which ecosystem you already use. Above that, buy a $200 or $500 tier only if your logs show you hit the $100 wall every week. The best AI chatbots are the ones whose meters match the way you work, and the only way to know that is to measure.
No vendor publishes a message count that allows a direct ranking. OpenAI says Pro 500 has the highest included usage among its Pro tiers, xAI lists Super at 15x Plus, Google lists Ultra at 5x AI Pro, and Microsoft says Microsoft 365 Pro has the highest usage among its consumer tiers. Each claim is relative to that vendor’s own plans.
Claude subscription limits work in five-hour windows. Anthropic raised those limits for Pro, Max, Team and seat-based Enterprise plans on 31 August 2026 and added a rate-limit reset that users can save and use themselves. Some premium Team seats also carry weekly limits.
Everyday text chats are unlimited, subject to abuse-prevention safeguards. File uploads, image generation, voice and data analysis each have separate limits, and the best models are not included. Free and Go users get standard access to GPT-5.6 Luna.
The confirmed heavy tiers cluster at about $100 per month. Claude Max starts from $100, Google AI Ultra was announced at $100, and Grok Super is $100. Perplexity Max is $200. OpenAI lists its Pro tiers at $100, $200 and $500 per month.
The figures in this article are US dollar list prices from vendor pages and news sources. We could not confirm euro prices for most plans. Expect euro pricing and VAT at checkout, and check the final amount before you pay.
Microsoft has not published a comparable numeric quota for Ireland. It describes Microsoft 365 Pro as having the highest usage for AI features among consumer tiers, and says long-running agent features and frontier models use usage-based billing. Copilot Business costs US$21 per user per month, paid yearly.
Wait for the window to reset, use a saved reset if your plan offers one, switch to a lighter model, or move urgent work to a pay-per-token API. If you hit limits every week, log the events for a month and compare them with the next tier up before you upgrade.
Niamh Kelly is the iGaming Editor at Tech Insider, where she previously worked as a freelance fashion journalist for The Irish News for three years and honed her media skills during her time at the BBC. At Tech Insider, she leads Ireland’s coverage with hands-on experience testing consumer and business technology, delivering in-depth analysis on AI, cybersecurity, cloud computing, and hardware trends shaping the future. Kelly was featured in RSVP online as part of their “Women of Style” series and has interviewed notable figures such as Katie Price and Louise Redknapp for major beauty product launches.
Tech Insider delivers in-depth coverage of the technologies shaping the future: AI, cybersecurity, cloud computing, hardware, and the trends that matter.
