AI with no message limit: only ChatGPT says unlimited on every plan, and it comes with an asterisk

Which AI has no message limit?

Five of 23 consumer plans call their use unlimited with no hedge word: ChatGPT Free, Go, Plus and Pro, and Perplexity Max at $167 a month. All five attach a policy clause. (Perplexity Free's "practically unlimited" covers basic searches only.) We read the pricing pages of seven AI chat vendors on 12 October 2026 and sorted every plan by what its page says about how much you can chat.

Not one of the 23 prints a number of chat messages. Seven plans, every Claude and Gemini tier, meter you on a clock that refills every five hours. The other eleven say "higher", "more", "generous", "extensive" or "limited", or name no limit at all. Here's the full read, in US prices, with each vendor's own words:

Sources are the vendors' own pages: ChatGPT pricing, Claude pricing, Google's Gemini limits page, Microsoft's AI credits page, xAI pricing, Perplexity's plan comparison and Mistral pricing. The vendors edit these pages, so treat every row as true on 12 October 2026.

What does unlimited mean on ChatGPT?

Unlimited on ChatGPT means everyday text chat has no fixed count, as long as OpenAI judges your use reasonable. The pricing page's asterisk reads "Usage must be reasonable and comply with our policies." OpenAI's help article on ChatGPT Go is blunter: "Usage is subject to plan limits, abuse guardrails, and temporary restrictions during high demand." So a ceiling exists. OpenAI just doesn't print it.

OpenAI announced the change on 6 August 2026. TechCrunch reported that day that "OpenAI is removing limits on text-based chats for all users on ChatGPT," with Free and Go accounts getting it the week after. The same report says OpenAI kept "separate limits for files, images, voice, and image generation." Those still stop you. A free account that attaches a PDF meets the plan card's "Limited messages with uploads".

The model is the second catch. ChatGPT's comparison table marks the flagship, GPT-6.1 Sol, as "No" on Free and Go and "Yes" on Plus. Free unlimited chat runs on the default model, not the best one. If ChatGPT tells you you've hit a limit anyway, the four ChatGPT limit messages tell you which wall it is.

Why do Claude and Gemini still cap messages?

Claude and Gemini meter compute rather than messages, and both refill on a five hour clock with a weekly ceiling above it. Anthropic's pricing FAQ says so directly: "Every plan has usage limits that reset on a rolling five-hour session window, and paid plans add weekly limits on top." It keeps a reserve right too: "we may limit your usage in other ways, such as weekly and monthly caps." Google uses the same clock: "Your limit refreshes every 5 hours until you reach your weekly limit."

Neither vendor counts messages, because a message isn't a fixed cost to them. Anthropic's help center says usage is "affected by several factors, including the length and complexity of your conversations," and it lists the model and the effort level as well. A long PDF in a Claude Opus chat spends more of the budget than a short question does. That's why nobody can tell you a number.

Paying more buys a multiplier. It doesn't buy an exit. Claude Max is "5x or 20x the usage of Pro per 5-hour session", and Google AI Ultra at $199.99 is 20 times AI Pro, yet both keep the weekly wall. Claude's reset rules and Gemini's six limits cover what to do when it lands.

Do Perplexity, Copilot, Grok or Le Chat say unlimited?

Only Perplexity does without a hedge, on its top plan, and its own pages argue with each other. Perplexity Max, at $167 a month billed yearly, sits under the heading "Unlimited usage and top performance", and its usage bullet reads "Highest usage limits and memory". Perplexity's help center goes further. Its plan comparison puts Max's Pro Searches on "Weekly limits (advanced use)". The Max footer adds "For personal use only and subject to our policies." Perplexity's free plan offers "Practically unlimited basic searches" beside 3 Pro Searches a day. Perplexity Pro limits breaks down the paid pools.

Microsoft promises "Extensive usage" of chat on Microsoft 365 Personal, Family and Premium, and its AI credits page, updated October 2026, gives a fourth tier, Microsoft 365 Pro, the same words. (Microsoft's US compare page carries no plan card for Pro, so it isn't in the table.) The same page explains why "extensive" isn't a count: "Feature limits are usage caps that help us ensure a good experience for all users." The free Copilot page names no limit at all, which isn't the same as having none.

Grok's pricing page promises free use "within generous limits", and SuperGrok at $30 gets "Higher rate limits across all features". Mistral sells Le Chat Pro at $14.99 with "More messages and web searches", then adds "Subject to fair usage limits and Mistral's Terms of Service." DeepSeek publishes no plan and no cap. Its terms, updated 27 March 2026, still keep the right to "add, upgrade, modify, suspend, or terminate services". How Grok's paid allowance refills is in Grok usage limits.

Is there any AI with truly no limit?

No major chat app offers one, and the nearest thing is paying per token. An API key from OpenAI, Anthropic or Google carries no monthly message count. You pay for every token in and out, and the ceilings are per minute rate limits rather than a monthly allowance. That's volume with a meter attached. A heavy month costs more, never less, and pay as you go AI compares the routes.

Smaller apps do advertise no cap, and they fill most of page one for this search. Shapes, a character chat platform, says it "doesn't meter your chats" on a page updated June 2026. InfiniChat AI's launch listing promises "unlimited AI conversations without limits". Neither page quotes a fair use clause or names the models behind the chat, so we couldn't check what stops them. Read the terms before you move your work there.

The last route is an open weights model on your own computer. Then the limit is your hardware, not anyone's policy. It's slower, and the best models need a serious GPU.

How many messages does a Whizi credit plan buy?

Whizi, which publishes this page, isn't unlimited either, and on raw volume ChatGPT Free beats every Whizi plan for $0. What Whizi adds is a printed count. Each plan carries a monthly credit budget, each model costs a fixed number of credits per message however long the message is, and every plan is capped at 10 messages a minute and 60 an hour. Credits reset when the UTC calendar month turns over, and unspent ones don't roll over. The credits page has every model's price, and as of October 2026 the plans buy this: