OpenAI is removing the usage cap on text-based ChatGPT conversations for free accounts. The Artificial Intelligence (AI) chatbot therefore becomes more useful for people who regularly edit text, answer questions, or plan tasks without paying for a subscription. At the same time, an Adobe integration expands its creative role, while research and media reports raise new concerns about trust and privacy.
What free text chats offer
According to the announcement of unlimited text chats, OpenAI is removing the previous usage limits for text-only conversations across all user accounts. GPT-5.6 Luna becomes the default model for Free and Go users, replacing GPT-5.5. A language model is an AI system that processes text and generates a suitable response.
“Unlimited” applies specifically to text chats. Uploaded files, images, voice interactions, and image generation will continue to have separate limits. If you want to analyze many documents or create numerous images, the change does not give you completely unrestricted free use.
OpenAI announced the update on August 6, 2026. The revised GPT-5.6 Sol became available to Plus and Pro subscribers that day, while the other changes for Free and Go accounts were due to arrive during the following week. Unlimited text chats and the new thinking feature were scheduled for the week after the announcement.
None of the supplied sources states the exact prices of the paid plans. The firm conclusion is therefore limited to standard text conversations being offered free of charge without a quantity cap. The reports also do not say whether Adobe functions, additional media use, or higher limits require other paid access.
What the new models change
GPT-5.6 Luna receives a “Think” button for harder questions. It is intended to allocate more reasoning effort to tasks that require several steps. Plus and Pro subscribers also get a slider in GPT-5.6 Sol that lets them adjust the amount of reasoning according to a task’s complexity and number of stages.
In OpenAI’s description of the model updates, the company says Sol should answer more directly, avoid unnecessary formatting, and adapt its level of detail to the question. A quick query should receive a concise answer, while planning, research, and writing tasks can receive fuller treatment. The model is also supposed to correct you when simple agreement would not be useful.
One example from the provider involves planning a bicycle ride after work. ChatGPT considers rain, wind, temperature, and coastal mist instead of supplying only a yes-or-no answer. The same style of structured response could support a multistep plan or help organize a research task at work.
OpenAI reports that an internal evaluation found factual errors to be 62 percent less common with Luna and 68 percent less common with Sol than with GPT-5.5 Instant. These figures come from the provider and were not independently verified in the supplied material. A lower measured error rate also does not make every answer reliable, so consequential claims still need checking.
The scale of the change is substantial: OpenAI says one billion people use ChatGPT each week. A separate analysis of worldwide ChatGPT use promises country-level insights, adoption trends, and changes in behavior. However, the available summary contains no specific country figures, so it does not support a firm conclusion about adoption in Switzerland.
What Adobe adds to the chat window
A new Adobe plugin, meaning an extension that adds functions to another service, brings more than 70 creative tools into ChatGPT, according to t3n. They come from applications including Photoshop, Premiere, Express, Acrobat, Firefly, Illustrator, and InDesign. The chat window is therefore moving beyond a text interface and toward a workspace for documents and visual material.
One concrete example is the creation of formatted PDF files from uploaded data and spreadsheets. Instead of merely summarizing the contents, ChatGPT can use Adobe functions to help with their visual presentation. Another example concerns an advertising campaign, where several images can be given a consistent style.
The integration of Adobe’s creative tools may reduce the need to switch between the chatbot and separate creative applications. A prompt, which is a written instruction given to the AI, could then cover work from an initial draft to document design. The sources do not state which functions are available without an Adobe subscription or what file limits apply.
Bundling these tools is convenient for occasional tasks. It does not automatically replace creative judgment: a consistent visual style may be technically polished yet unsuitable for its purpose. For business documents, you still need to check whether content, figures, and brand details were transferred correctly.
Pros and Cons of ChatGPT as an everyday tool
Broader access lowers the practical barrier to longer projects, but it also magnifies familiar weaknesses. A particular concern is that a natural-sounding conversation can inspire more trust than the system’s technical reliability warrants.
Pros:
- More room for free use – Longer text-based conversations will no longer be interrupted by the previous usage cap.
- Adjustable reasoning – The Think button and the slider for paid accounts provide more processing for complex tasks.
- Broader creative work – Adobe tools connect writing, PDF preparation, and image editing within a conversation.
- More focused answers – According to the provider, GPT-5.6 Sol should be more concise while retaining enough context for larger tasks.
Cons:
- Limited “unlimited” access – Files, images, voice, and image generation remain subject to their own caps.
- Unverified reliability – The reported reductions in factual errors are based on OpenAI’s internal evaluation.
- Confidential input – Natural conversation can encourage people to share personal or intimate information with a service run by a US company.
- Distorted trust – People’s assessment of a text can depend on whether they believe a human or AI produced it.
A study involving more than 2,500 participants demonstrates this trust problem particularly clearly. The participants could not distinguish human short stories from ChatGPT-generated fiction more accurately than chance. In the first experiment, which involved 1,682 people, the AI stories received higher ratings for perceived quality and immersion, meaning the feeling of being absorbed in a story.
Labels changed the assessment, however. Regardless of a story’s real origin, participants rated it more highly when they were told a human had written it. The findings on AI-generated short stories therefore do not simply demonstrate superior creative quality. They show that content, assumed authorship, and personal attitudes toward AI all affect the judgment.
Privacy creates another concern. A report on personal conversations with chatbots argues that a private-looking chat window encourages people to disclose intimate information and forget that they are not speaking with another person. Under certain conditions, responses tailored to the conversation can also be persuasive in political discussions and influence opinions, according to the report.
What this means for you
If you are a beginner, a sensible first move is to choose a clearly limited task that contains no confidential information. You could improve the structure of an existing text, clarify a general question, or request a multistep plan. This lets you learn how the conversation works without immediately disclosing personal, professional, or intimate details.
Step 1: Begin with a manageable task
- Describe the desired result and its intended purpose in a few sentences.
- Leave out names, internal documents, and personal details unless they are necessary for the task.
- For a complex question, use the Think button to request more thorough processing.
Step 2: Review content and design separately
- Check factual statements, figures, and conclusions against the material available to you.
- For Adobe outputs, also review readability, visual style, and whether the content was transferred correctly.
- Identify AI-generated material transparently when its origin is relevant to other people.
If you are an advanced user, you can gain more by matching the reasoning effort to the task and dividing large projects into verifiable stages. For research, you might first structure the issue, then process individual points, and finally ask the system to identify contradictions. Creative work benefits from the same separation between content, visual execution, and final human judgment.
For Switzerland, the main point so far is that OpenAI describes unlimited text chats as a general update for free accounts. The sources provide no separate details about Swiss availability, supported national languages, or the processing of data from Switzerland. OpenAI’s country-level analysis also supplies no Swiss figures in the summary provided.
Unlimited free text conversations and Adobe integration make ChatGPT a more practical tool for planning, writing, and design. The benefit is growing, but so is the amount of information people may entrust to a convincingly worded system. The unresolved risk lies less in a single wrong answer than in becoming accustomed to treating fluent language as evidence of knowledge, confidentiality, or human judgment.
Sources
- ChatGPT brings unlimited text chats to free users – TechCrunch, 2026-08-06
- Improving GPT‑5.6 Sol in ChatGPT—and expanding access to GPT-5.6 Luna for free users – OpenAI, 2026-08-06
- Über 70 Kreativ-Tools integriert: Adobe-Plugin macht ChatGPT zur Canva-Alternative – t3n, 2026-08-07
- From asking to doing: How the world is putting ChatGPT to work – OpenAI, 2026-08-06
- Leser bewerten ChatGPT-Geschichten höher als menschliche, bis sie die KI-Herkunft erfahren – The Decoder, 2026-08-08
- Mehr als nur ein Chatfenster: Wir vertrauen KI viel über uns an – werden wir so manipulierbar? – t3n, 2026-08-07
