ChatGPT opens unlimited text chats to free users

ChatGPT takes a new step toward making everyday use free

OpenAI plans to allow free ChatGPT users, as well as subscribers to the Go plan, to exchange messages without a cap in text conversations. The information was reported by TechCrunch in its article entitled “ChatGPT brings unlimited text chats to free users”. According to the outlet, the rollout is set to begin as early as next week.

The wording matters: this is not a promise of unlimited access to all of ChatGPT’s capabilities, but an opening up of text chats. The distinction sheds light on OpenAI’s choice. Simple written exchanges, which for much of the public are the entry point to a conversational assistant, are becoming a more generous loss-leader offering. By contrast, functions requiring more computation, multimodal tools, or capabilities reserved for paid plans remain central to the company’s commercial differentiation.

This development comes in a market where free access is never truly uniform. Generative AI assistants frequently offer a trial experience or access without payment, but this is often governed by usage limits, temporary restrictions, models available depending on the plan, or caps that can change with demand. For users, these rules can sometimes be difficult to understand: a conversation may work normally, then be interrupted because a quota has been reached or because a given feature is no longer accessible until after a certain period.

By removing the cap on text for the free tier, OpenAI is therefore seeking to reduce a very concrete friction point. A user should no longer have to count their messages when using ChatGPT to rewrite an email, prepare an outline, ask general knowledge questions, translate a passage, organize a task, or continue a work discussion. This development has less to do with the technological demonstration effect than with frequency of use: the goal is to make the assistant available as a daily interface, rather than as a service to be used sparingly.

The choice of text is also consistent with the product’s recent history. ChatGPT was publicly launched by OpenAI in November 2022 and established itself as one of the most visible faces of generative AI among the general public. Its conversational format helped popularize the idea that a language model could serve as a general-purpose tool: writing, explaining, summarizing, programming, brainstorming, or assisting with processes. Since then, expectations have shifted. The issue is no longer simply discovering conversational AI, but deciding which assistant becomes the default reflex.

In this context, a message cap is more than a technical or pricing constraint. It can prevent a habit from forming. A user who knows they risk losing access after a few requests will reserve the tool for certain moments. Someone who can exchange messages freely is more likely to centralize ordinary tasks there. For OpenAI, this difference is strategic: the value of a conversational platform also depends on its place in routines, familiarity with its interface, and users’ propensity to return without having to systematically compare alternatives.

The Go plan is explicitly included in this change, according to TechCrunch. This detail shows that OpenAI is not presenting the opening up of text chats as an exclusive privilege of the most expensive plans. Instead, the company appears to be shifting the boundary between the basic layer, now more accessible, and the experiences that justify payment: more power, speed, features, or capabilities with higher infrastructure costs. Unlimited text thus becomes a baseline use case, while the business model must play out elsewhere.

The announced facts: uncapped written exchanges and a Think button

The core of the announcement is simple: free users and those on the Go plan will be able to have uncapped text conversations. TechCrunch indicates that the change is expected as early as the following week. OpenAI is therefore not merely describing a raised limit or a temporary quota increase: the outlet specifically refers to unlimited text chats for these two categories of users.

However, the exact scope of this promise must be retained. The term “unlimited” here applies to text conversations. On its own, it does not allow one to conclude that all interactions with ChatGPT, regardless of the model used, the level of reasoning requested, access to files, images, voice, or other tools, will also be unrestricted. In AI products, cost and availability vary significantly depending on the type of task. The scope reported by TechCrunch should therefore be understood for what it is: a major easing of ordinary written use.

OpenAI is also set to introduce a button called Think. According to the reported details, this control is intended to allow users to trigger more reasoning when a query is complex. The word “trigger” is essential: users will have a more explicit means of indicating that their request deserves deeper analysis than an immediate conversational response.

This button could address a classic problem with generative assistants. The same interface is used for very different requests: correcting a sentence, producing a list of ideas, interpreting a dense text, solving a logic problem, analyzing conflicting constraints, or structuring a line of reasoning. Yet not all of these uses involve the same level of computation or processing time. A very short interaction does not necessarily call for the same resources as a question requiring several stages of thought.

With Think, OpenAI is making this distinction more visible in the product. Rather than leaving users to guess what behavior the assistant will adopt, the interface is intended to offer them a simple signal: for certain questions, it is possible to request more reasoning. TechCrunch does not detail the exact parameters of this feature, its possible cost, the limits that will be associated with it, or the models concerned. It would therefore be premature to infer universal and unconditional access to advanced capabilities.

But the arrival of such a button reveals a clear direction. Consumer conversational AI is no longer limited to a single category of response. Publishers are seeking to offer a gradation between rapid assistance and deeper processing, between everyday tasks and complex problems. This differentiation may be technical, but it is also educational: it helps the public understand that an AI can be used in several ways and that not all queries have the same profile.

For OpenAI, the combination of the two announcements is particularly coherent. On the one hand, ordinary text exchanges become much more accessible. On the other, a control makes it possible to indicate that a case requires more reasoning. The product thus more clearly separates the basic conversational flow from requests that may be more demanding. This architecture is compatible with an economic logic in which free access serves to maximize the service’s presence, while the most costly resources are managed with greater control.

The communication reported by TechCrunch provides no details on the operational conditions of the rollout. As with many changes to digital products, access may depend on the timeline for availability in apps and on the web. For users, the point to watch will therefore be less the general wording of the announcement than its concrete implementation in their interface: the actual disappearance of the text-message cap, the presence of Think, and a clear indication of the limits that would continue to apply to other types of use.

The economic bet: making free access a distribution tool

OpenAI’s decision can be read as a distribution bet. An AI assistant does not win only through its performance in a one-off demonstration. It also wins when it becomes the tool that people open spontaneously, before a search engine, word processor, code editor, or another application. To reach that status, frequency matters as much as perceived quality. Removing a cap on text exchanges reduces one of the main reasons not to use the tool continuously.

This strategy is all the more significant because the generative AI market has become highly competitive. ChatGPT faces assistants offered by groups that already have vast product ecosystems, as well as players specializing in language models. Google, Microsoft, Anthropic, Meta, and other companies occupy different positions: some have search platforms, operating systems, office suites, social networks, or cloud infrastructures. In such an environment, free availability is a lever for maintaining a direct relationship with users.

The issue is not only acquiring new accounts. It is about making ChatGPT a habit before that habit is captured by a competing interface. A user who has built their reference points, working methods, and trust around an assistant does not necessarily switch with every new model announcement. They may compare, but they also have to relearn an interface, adjust their prompts, and assess the consistency of responses. Reducing usage friction therefore creates an advantage that is not solely technological.

This logic recalls a common principle in digital platforms: the free layer can serve to embed the product in everyday life, while monetization rests on users who need more guarantees, power, or features. In generative AI, this separation is even more structuring because costs are not identical across uses. Answering a short message and carrying out more demanding reasoning do not necessarily draw on the same resources. Processing rich content or providing advanced capabilities can also represent a greater burden.

Unlimited text therefore enables OpenAI to offer an easily understandable promise: it becomes possible to talk to ChatGPT without watching a counter for written interactions. This is an immediately tangible benefit, particularly for people who do not want to choose a subscription before integrating the tool into their uses. At the same time, the distinction between basic use and more costly functions can preserve the appeal of paid plans. Free access does not necessarily eliminate monetization; on the contrary, it can broaden the top of the journey that leads, for some users, toward premium services.

The Think button fits into this same architecture. It can make more visible the fact that certain queries require additional effort from the system. This visibility is useful both for the product and its economics. It avoids presenting all responses as equivalent and makes it possible to direct users toward a mode suited to the complexity of their request. TechCrunch does not specify the access policy associated with Think, but the very existence of this control suggests that the interface is seeking to better distinguish types of interaction.

OpenAI must nevertheless manage a classic tension. The more generous the free offering, the more it promotes adoption, retention, and awareness. But it can also increase infrastructure demand and reduce the immediate incentive to pay. The balance depends on the company’s ability to reserve sufficiently useful advantages for commercial offerings without making the free product too frustrating. Removing the cap on text appears to be aimed precisely at this compromise: substantially improving the ordinary experience without claiming that all intensive or advanced uses become indiscriminately free.

This approach can also be understood as a response to the gradual normalization of AI chat. When a user sees an assistant as a basic tool, a message limitation appears harder to accept than when the technology was still experimental. By raising the service level of the free offering, OpenAI is anticipating this increase in expectations. The company is no longer only selling access to something new; it is seeking to maintain preference for its interface in a landscape where many players now offer conversational experiences.

The economic scope of the announcement will ultimately depend on its stability. A promise of uncapped text conversations has particular force if it is durable and understandable. If users clearly understand what is freely accessible, what belongs to an advanced tier, and what requires a subscription, the separation between free and paid becomes simpler. Conversely, fluctuating or opaque rules could limit the sought-after distribution advantage. For OpenAI, the challenge is therefore not only to lift a limit: it is to establish a product hierarchy that the public can truly understand.

A signal for the European and French-speaking market for AI tools

In France and more broadly in French-speaking markets, the announcement may have a very concrete effect on the daily adoption of AI. Price remains a barrier for a significant share of the public, particularly among students, freelancers, associations, small businesses, and employees who do not have a subscription funded by their organization. Uncapped text access lowers the barrier to prolonged experimentation: it becomes easier to test an assistant’s usefulness over time without having to weigh each conversation.

For individual uses, the most immediate applications are often text-based. They include writing assistance, summarizing notes, preparing for an interview, rewording a message, explaining a concept, creating an outline, or language assistance. These needs do not always require multimodal functions or complex processing. This is precisely why the opening announced by OpenAI may be more consequential than an improvement reserved for a narrow category of expert users: it affects the foundation of the most frequent uses.

In French organizations, this development could also strengthen the presence of consumer AI tools in informal practices. This is both an opportunity and a point requiring vigilance. Easier access can speed up adoption, but it does not exempt employers, public administrations, or educational institutions from setting rules on the data that can be submitted to an external assistant. The fact that a service is free does not change the requirements for confidentiality, protection of sensitive information, or internal compliance.

The European context adds a particular layer to this adoption. Users and organizations on the continent are attentive to data management, tool transparency, and the regulatory framework applicable to AI systems. Improved access to ChatGPT does not settle any of these questions. On the contrary, it may make them more prominent: the more the tool is used daily, the more important the choices related to shared information, professional rules, and user training become.

The Think button may also have educational value. Many people are still discovering that the quality of a response depends on how the request is framed, the context provided, and the expected level of analysis. A control that encourages users to request more reasoning may help clarify this difference. It does not guarantee the accuracy of responses and does not replace human verification, particularly for high-stakes legal, medical, financial, scientific, or professional matters. But it may encourage less superficial use of the assistant for complex requests.

For software publishers, training platforms, and French integrators, broader free text access to ChatGPT is also a competitive factor. If a large part of the public can easily access a general-purpose assistant, local or specialized products must demonstrate distinct value: integration into business tools, command of specific domains, hosting, support, security, compliance, or the quality of the French-language experience. Free access does not make these needs disappear, but it raises the level of comparison that new services face.

Competition will also play out through distribution. Major technology players have different channels for reaching users: mobile apps, search engines, productivity suites, operating systems, or cloud services. ChatGPT was built as a standalone destination, which gives the removal of the text cap particular value. OpenAI is seeking to retain the centrality of that destination by making its interface a place users can return to without wondering whether they still have enough messages available.

For the French-speaking market, the issue of language should not be reduced to translation alone. Users expect assistants to understand registers, writing standards, administrative references, or local professional contexts. Broadening free use can multiply opportunities to assess this quality in practice. It can also make differences between services more visible, especially when it comes to producing professional texts, explaining complex concepts, or preserving nuances in a long conversation.

Toward two-speed conversational AI, between universal access and premium compute

The announcement reported by TechCrunch outlines a likely trajectory for generative AI products: basic text chat is tending to become a commodity, while commercial value shifts toward capabilities that require more resources or provide a more directly measurable advantage. From this perspective, opening up written conversations is not an abandonment of the paid strategy. It is a clearer redefinition of what must be free in order to remain competitive.

This development could change the very perception of ChatGPT. When a service is subject to visible quotas, it may be seen as a scarce resource to preserve for specific needs. When it becomes uncapped for text, it moves closer to being a permanent support tool. This availability encourages repetitive use, longer exchanges, and exploratory requests. It may also shift expectations: users could consider it normal to have an accessible conversational assistant at any time, much like a search engine or messaging service.

The question then becomes where the boundary lies between ordinary assistance and enhanced reasoning. Think provides a clue as to how OpenAI wishes to organize that boundary. The control implicitly acknowledges that a rapid response is not always enough and that users must be able to signal a more complex request. In the long term, AI interfaces could become more explicit about the trade-offs between speed, depth, cost, and perceived reliability. Such transparency would be preferable to an experience in which the public completely ignores why certain responses change behavior or become less accessible.

However, requesting more reasoning should not be confused with an absolute guarantee of quality. Language models can produce convincing responses that still require verification, especially when the consequences of an error are significant. The arrival of Think does not remove this fundamental rule. It may improve how the system approaches certain tasks, but it should not encourage blind delegation of judgment to the tool.

The challenge for OpenAI will be to preserve a sufficiently smooth free experience while preventing broader access from weighing on service quality. Users will judge the announcement on very practical elements: the actual availability of text chats, continuity of conversations, understanding of the remaining limits, and the concrete usefulness of the Think button. The promise of unlimited free access is powerful because it is simple; it will therefore create high expectations for product consistency.

For competitors, the pressure is less tied to an isolated feature than to the change in benchmark it may establish. If the public gets used to using ChatGPT without a cap on text messages, basic limits imposed elsewhere will become more visible. Other assistants may respond by improving their own offerings, through deeper integrations into their ecosystems, or by differentiating themselves through specific functions. The market will not be limited to free access: quality, trust, integration, and specialization will remain decisive. But the minimum expected level for conversational access could rise.

In France as in Europe, this competition may accelerate the normalization of AI in everyday tools. It will also require users to distinguish more carefully between a free general-purpose service, advanced paid functions, and solutions designed for professional or regulated needs. As unlimited text becomes a possible standard, the debate will shift toward the conditions under which complex requests are processed, access to computing resources, and companies’ ability to offer AI that is at once useful, understandable, and economically sustainable.

OpenAI’s bet is therefore clear: to broaden ChatGPT’s entry point as much as possible, so that the assistant remains a daily reflex in an increasingly crowded market. The next step will be to observe whether this free text access is enough to durably strengthen that position, or whether users will ultimately favor assistants that are best integrated into their tools, data, and working environments. It is at the intersection of mass audience and premium capabilities that a significant part of the competition ahead will play out.