In case you haven’t heard, there are now several free AI programs available that let you email, summarise a webpage, create graphic images, translate a piece of text, or create a budget without having to spend any money. Many users find it difficult to keep track of which applications stop users at which points because the free versions of the various services count usage in different ways. Limits on usage were changed many times over the course of 2026. Therefore, if you have seen a guide to the free AI applications in the spring of 2026, that guide may not be current.
The most significant change to ChatGPT’s free version is that unlimited plain text chat has been added. As a result, many users have assumed that every single feature on ChatGPT was also unlimited. However, that assumption is incorrect. All the information provided below regarding the limitations of the free version was verified on the companies’ respective support sites and within their version tables on 29 September 2026. Each of the companies will frequently update their support sites so the values may not match the limit set by their website. After you understand the differences between what counts towards usage, you can choose to use only those tools that will not stop you in the middle of a project.
The 2026 Free-Tier Reality: Message Caps, Credits, and Context Windows
Late in 2026, the term “free tier” does not accurately describe limiting factors on platforms. Each type of limiting factor controlled by a service provider varies, and so trying to use message counts to provide a direct comparison is misleading.
ChatGPT removed the message cap on text chat. Since 06/08/2026, text chats have been uncapped on every ChatGPT plan, and free accounts run on a model called GPT-5.6 Luna. OpenAI’s own announcement said separate limits remain for files, images, voice and image generation.
Gemini measures effort rather than messages. Google says its limits are compute-based and “factor in the complexity of your prompt, the models and features you use, and the length of your chat”. The allowance refreshes every 5 hours until you reach a weekly limit. In plain terms, a long document or a picture uses up more of your allowance than a short question does.
Claude is designed to operate under time schedules. Anthropic does not publish the message limitation for its free plan, but notes the amount of use will be determined based on the length and complexity of your conversations, the features you utilize, and the model used. The Pro Edition of Claude allows you to use at least five times more than the Free Edition per session, with the session reset after five hours of inactivity.
Perplexity restricts its advanced queries. The table on its website indicates a limit of three Pro Searches per day and a limit of one Research Query per month on its Free Edition, with unlimited basic queries. Some sources continue to state there are five Pro Searches per day, but Perplexity now states three in its own table.
The final limit to consider is the context window. This is the maximum amount of text that AI can process in one pass, and it determines if a lengthy document fits within AI’s context window. For example, Google provides the following context window limits for its Gemini product: 32K tokens without a paid account, 128K tokens with an AI Plus plan, and 1 million tokens with an AI Pro plan. A token represents a small part of a word.
| Tool (free plan) | What it counts | Published free allowance, as of 29/09/2026 | When it comes back |
|---|---|---|---|
| ChatGPT | Files, images, voice (text is uncapped) | No numbers published for files or images; 500MB Library storage | Rolling windows that start from first use, not midnight |
| Claude | Overall usage per session | No number published | Five-hour session |
| Gemini | Compute (effort per prompt) | “Standard limits”; 32k-token context window | Every 5 hours, inside a weekly cap |
| Perplexity | Pro Searches and Research | 3 Pro Searches/day, 1 Research/month | Daily / monthly |
| Microsoft Copilot | Chat, images, uploads | Microsoft says only “subject to available capacity and limits” | Microsoft doesn’t say |
The privacy settings for these AI technologies differ. For example, with ChatGPT, go to the settings page, then click on Data controls and uncheck the box next to “Improve the model for everyone.” Once this has been done, your new conversations then won’t be used to train its models. By contrast, Claude provides this option in its settings under Privacy. Lastly, turning off “Keep Activity” in Gemini will prevent future chats from being used for training, however, Google will retain these records for a maximum duration of 72 hours.
Writing and Planning: Navigating Text and Chat Limits
Mail and Budget types (emails and budget plans) are primarily typing applications, and they offer their “free” plans with generous typing limits. The typing limits become an issue when you upload files or request more advanced features through that same instance or conversation.
ChatGPT is the easiest free option for long writing sessions right now, because plain text has no message count. OpenAI lists the costs. In the United States, free users have seen labelled ads under some answers since 09/02/2026. Whether those ads have reached Nigeria couldn’t be confirmed for this article. Free users also get a Think button that gives the model more time on a hard question, like checking a monthly budget. OpenAI puts guardrails against abuse on heavy use of it. The newest model, GPT-6 Astra, isn’t available on Free or Go at all, according to a tracker citing OpenAI’s help pages.
Claude operates under a different model. Each word and character in a message uses a portion of the total five-hour usage based limit. Long pastes are more resource-consuming, meaning that Anthropic claims that users can better manage their usage limits by only starting new chats once they are approaching the end of their five-hour usage period. Longer conversations that trigger automatic context management consume more of your usage. Having done so, Anthropic recommends that users create one long prompt for drafting their cover letters and then go back and re-write them 10 times as needed; it also recommends grouping related requests into one message when drafting responses. In its Help Centre, Anthropic explicitly provides “Usage Progress” bars located beneath the Settings menu to show users how much of their usage is remaining.
When using these free tools, students should be aware of watermarking. In August 2026, TechCrunch reported anthropic’s decision to add a watermark to the text or files created by Claude, in order to be compliant with EU regulations. Additionally, OpenAI’s terms clearly state that you cannot represent that the output was created by a human when it was not. A chatbot can be beneficial when you want to organize and revise an essay, however, representing the chatbot’s written material as your own would violate the rules established by the provider of the chatbot. There are currently some watermarks implemented for detecting such behavior.
A budget and all related monetary tasks also come with a caveat. OpenAI states that you should not rely on the generated data from AI as an alternative to that of a professional financial expert, and you should not rely on output as a substitute for professional advice. All numerical values generated by AI should be verified to your own personal or business financial statements, prior to any actions being taken based upon them.
Summarising and Translating: File-Size Walls and Character Counts
The limitations for summarising a paragraph in each tool’s format are different. The limitations are based on the number of times you can upload a file and how many tokens each tool has.
The limitation on long documents when using Gemini’s free plan is the 32k-token context window size. A very short report would easily qualify, but an entire chapter from a textbook will be unlikely to meet the requirements of being summarized. The number displayed by Gemini does not decrease as the original file is summarized. Google has said only that when media is generated or a deep research is done, those usages consume more of your available usage; and that uploaded files count as “complexity” in the same manner as deep research and media generation.
Perplexity does an excellent job of summarizing a web page due to its citation of the source of the information, however, the free plan limits the user to only [basic file uploads, marked “Limited”] for files, and it does not show you a total number of uploads used. According to one person who reviewed aisparkhub’s testing of Perplexity on the free plan back on 09/02/2026, there were times when [Perplexity sometimes misunderstood uploaded material] and they needed to verify that the answers provided were in fact accurate by reviewing the original sources themselves. While this is only one individual experience with Perplexity, and does not provide an average error rate, you may want to emulate their habit of verifying the cited sources of any summaries before trusting them.
ChatGPT users can only upload a certain number of files, separate and apart from what is included in their text chats for free accounts. The limited files uploaded for each user also have a size restriction and count against their total Library size of 500MB for free accounts.
DeepL has the most clear free limits regarding translation. It lists as part of its plan comparison for their free version, a limit of 50,000 characters a month and 1 file translation of up to 5 MB that can be translated. Some sites that are third-party to DeepL offer a figure of 500,000 characters, however, this is for DeepL’s separate Free API Plan used by developers; they cannot be accessed through the translator that you use via your browser. According to one writer for eesel.ai, 50,000 characters roughly equates to about 8,300 words a month. This is enough to write a few letters, but would leave you with little to no margin for a thesis type of work. If you have one long document that you need to translate, then only one file can be translated in a single month with a DeepL free account, so it would be wise to break up the text that you are going to translate into smaller sections to get the most out of your free allowance.
Generating Images: Daily Boosts, Watermarks, and Usage Rights
Images consume far more than any other voluntary allowance and have the least openly available limits.
Microsoft Copilot is typically the most recommended tool for finding free images as the ‘boosts’ (credit) of fast image generation are available to free users. Articles published prior to Microsoft’s App Update (like copilot-experts.com in July 2026) indicated free users would receive an allowance of 15 image generation boosts per day through Microsoft’s Copilot. On August 18, 2026 Microsoft moved to the new version of the app and now only states on the Microsoft website that you can create images for free, “subject to available capacity and limits”. The website further states that if you exceed your limit you can upgrade the plan or return at a later date to continue. Thus, the figure of 15 boosts is not guaranteed but can be treated as being associated with the previous version.
As a result of the lack of visible numbers there is a frustration among users. A Copilot user on Microsoft Q&A stated they [kept getting the image-limit message but couldn’t find out what the limit was or when it would reset] without being able to find out when the limit would be reset. A Microsoft Q&A response only described the system as being dynamic and recommended that users wait at least 24 hours from their last image to generate another.
The Gemini application provides users with a free plan of providing Image Generation and Media Generation, with each having separate limits on both text and images. As a result, users can spend more on Image generation during the day than on text generation, as media generation consumes more of your usage. ChatGPT Free Plan limits Image Generation similar to Text Generation on per account basis, keeping its own cap on image creation. Perplexity Free Plan does not include Image Generation.
Watermarks: Google announced on August 12, 2026, that users will be able to remove the visible “Gemini” watermark from their photos through their settings (Settings > Media Watermark). That feature will be rolled out in the following days. However, the invisible SynthID watermark and C2PA metadata will still remain in the images. The C2PA metadata is hidden in the image file and contains information on where the image originated. OpenAI states that images generated with ChatGPT contain both C2PA metadata and SynthID watermarks. Removing the visible watermark does not make an image completely untraceable.
When it comes to using images for commercial purposes, having appropriate usage rights is critical if you plan on selling a product. OpenAI’s Agreement clearly states that Output belongs solely to User when created using OpenAI’s services, you own the output. OpenAI has also made it clear through its Copilot Consumer Agreement that User is only permitted to utilize Copilot for Personal Use and that User shall not delete or alter the Content Credentialing System which identifies that Copilot created a specific image or other Content using Copilot as the Creation Source. Therefore, free images produced by Copilot would be a poor choice for business flyers, but an acceptable choice for birthday cards. This is not legal advice. To ensure compliance, please review OpenAI’s Terms of Use before implementing images created by OpenAI in a professional manner.
The “No Sign-Up” Route: When You Just Need One Quick Answer
There are times when creating an additional account may not be worthwhile for particular functions or operations. Even so, there are still a few methods to access artificial intelligence without needing an account. The currently available methods are more limited than suggested by most reference material.
The most documented resource is Duck.ai, a creation of DuckDuckGo (DDG). The information on the DDG website indicates that no account is required for any user; DDG also states it doesn’t record or store your chats, and they aren’t used to train models by DDG or the model providers behind it. Each user has a ceiling on the number of times they may use the DDG free version per day. By default, this ceiling resets itself at midnight UTC. In Nigeria, which has a West Africa Time (WAT) UTC/GMT+1 difference, this reset will take place at one o’clock local time. Further, DuckDuckGo advises users that they may see an error when using a VPN or a cellular network and suggests that users switch to a Wi-Fi connection.
ChatGPT also works without signing in. OpenAI’s help pages describe signed-out use, where your training setting is saved in that browser. You lose saved history and the account features. Gemini supports signed-out use as well: Google says such data is handled under its general privacy policy and not linked to a Google account. A 2025 9to5Google report added that uploading files still required signing in and that the Android app still needed a login.
The number of guest routes available has continued to decrease. One requirement of Copilot is that you have a Microsoft Account in order to use it. Recently, a member of Microsoft’s Tech Community reported that Edge in Private (InPrivate)/Android now requires a sign-in for guest access that does not allow for skip options. Perplexity continues to be advertised as an option for basic search while logged out. On Reddit in March 2026, however, one user shared how they were eventually no longer able to access any of the answers that they had previously found through their login overlay on their browser, which eventually blocked every answer. Although these are just user experiences that have been reported and not necessarily indicative of any current official policy, they do provide some potential evidence that the guest route can close unexpectedly.
Decision Tree: Match Your Task to the Right Free Tier
First, perform the task you have to complete in 10 minutes, then follow the branching points. The editorial summary provided is only a summary of the categories listed above. Each category has a reason to be omitted from consideration by every vendor.
- Do you need to avoid creating an account?
- Yes, and privacy matters most: use Duck.ai. It’s good for a quick rewrite or question with no chat storage by default. The downsides are an unpublished daily cap that resets at 1:00 a.m. WAT, and possible errors on mobile data.
- Yes, but you want a familiar chatbot: use ChatGPT signed out. You get quick answers, but no saved history and fewer features.
- No: go to question 2.
- Is the task mostly typing, such as an email, CV, caption or budget outline?
- Use ChatGPT Free. Its plain-text chat is uncapped, so long drafting sessions won’t hit a wall. The downsides are the ad programme in some markets and no access to OpenAI’s newest model.
- Choose Claude if you prefer its writing style and can batch your requests. The downside is a five-hour allowance you can use up quickly with long pastes.
- Is the task about a web page or a current fact?
- Use Perplexity’s basic search, which comes with cited sources. The downside is only 3 Pro Searches a day, and you still need to check the citations yourself.
- Is the task a document you want to upload?
- For a short file, Gemini or ChatGPT will do. Gemini’s downside is its 32k-token reading capacity, and ChatGPT’s is an unpublished file cap.
- For a very long file, split it into sections before you upload. No free plan here publishes a limit that guarantees a long document will fit.
- Is the task a translation?
- Use DeepL for a paragraph or a letter. The downside is 50,000 characters and one 5 MB file a month.
- Is the task a picture?
- For a personal image, use Copilot or Gemini. Copilot’s downsides are that it doesn’t publish its limit and that its terms say personal use only. Gemini’s downside is that images use up your chat allowance too.
- For anything you’ll sell, use ChatGPT, whose terms assign the output to you. Its free image cap is separate and not published.
Selecting your tool is only part of what you need to do to achieve your goal. Providing a clear first request will utilize fewer of your allowances than any other option, and this will be critical when using either Claude or Gemini because every communication you make counts towards your maximum.
What to Do First
You can complete your five daily tasks without any charges in September 2026 if you use the appropriate tool based on its limitations: ChatGPT for longer documents; DeepL for shorter translations; Perplexity for cited sources, and either Copilot or Gemini for personal images. Before pasting your personal information into these tools, make sure you have visited the tool’s privacy settings and disabled the model training feature. According to the suppliers of these tools – namely, OpenAI, Anthropic, and Google – by doing this now, you will prevent your personal information from being included in any future training datasets, and will help ensure that you do not encounter an unexpected limit with regard to a task that is time-sensitive. If you do not do so, you risk having an incomplete CV at midnight, and could inadvertently share your personal data with others who do not need it.
Frequently Asked Questions
What time do free AI limits reset in Nigeria?
Duck.ai restarts every day at midnight UTC, which corresponds to 1:00 AM in West Africa (WAT), whereas Claude and Gemini are refilled on a rolling basis every five hours, while ChatGPT’s cap resets to zero after the first time it is reached, and does not reset at midnight, start counting from your first use.
Why does Duck.ai give me an error on mobile data?
DuckDuckGo indicates that their anonymous way of applying usage limits may not work properly while using a VPN or via cell phone, which can trigger errors on VPNs or cellular networks. To resolve this issue, switch to Wi-Fi when you are experiencing problems and report the issue via the Share Feedback button.
Can I use a free Copilot image on my shop flyer?
Probably not; according to Copilot’s consumer terms of agreement, Copilot can be used only for your own personal use; however, ChatGPT allows you ownership of your outputs based on the current agreement.
Does turning off AI training delete my old chats?
No; according to OpenAI, disabling “Improve model for everyone” does not remove or hide saved chat histories. You still need to manually delete your chat history in order to remove it.
