The biggest ChatGPT update this week is not another model name. It is the removal of a stop sign.
OpenAI says Free and Go users will get unlimited text chats starting next week, subject to abuse guardrails. Their default model is moving to GPT-5.6 Luna, and a new Think button will let Luna spend more effort on harder questions. Files, images and other tools will still have separate limits.
Free and Go users can keep text conversations going without the old text rate limit once the rollout reaches them, while tool limits remain.
UAE residents who use ChatGPT for work, study, research, planning, writing or everyday questions.
Longer text-only workflows become possible without upgrading simply to avoid a message cap.
Unlimited applies to text chats, not files, images or other tools, and Think uses Luna rather than the more capable Sol model.
Luna becomes the Free and Go default this week; unlimited text chats and Think begin rolling out next week.
If text is your main use, try the new Free experience before paying solely for more messages; pay when you need stronger reasoning or higher tool limits.
Free users gain the clearest upgrade, while Go users also receive unlimited text and Luna-based Think access.
Free and Go users in supported ChatGPT countries, including the UAE, as the rollout reaches eligible accounts.
Watch the rollout on your account, remaining tool limits, and any clarification to OpenAI’s paid-model routing documentation.
For UAE users, this changes a basic question: when do you actually need to pay? If most of your ChatGPT use is typing, planning, studying, writing or asking follow-up questions, the Free tier is about to become much harder to outgrow. But unlimited text is not unlimited ChatGPT, and Luna is not the same model paid users get.
What Actually Changes
OpenAI announced the update on August 6. GPT-5.6 Luna is becoming the default model for Free and Go users this week. Starting next week, those users are due to receive unlimited text chats and a Think button for questions that need more reasoning.
The word “unlimited” needs the qualifier beside it. OpenAI says the change is for text chats and remains subject to abuse guardrails. File uploads, images and other tools keep separate limits. So a long conversation made only of text can keep going, while a workflow built around PDFs, screenshots or image generation can still hit a ceiling.
That difference matters because text is the part most people touch first. The old Free experience trained users to think about conserving messages. Once that barrier disappears, ChatGPT starts to feel less like a metered demo and more like a normal utility.
| Plan | What changes | What still matters |
| Free | Luna default; unlimited text and Think rolling out next week. | Files, images and other tools stay limited; Think uses Luna. |
| Go | Same Luna text change and Think access. | Go still offers expanded access to tools such as uploads, images and data analysis, plus longer memory. |
| Plus | Updated GPT-5.6 Sol access and reasoning slider. | Reasoning limits still depend on plan; OpenAI documentation is not fully aligned on the Instant label. |
| Pro | Updated Sol experience, slider and higher reasoning options. | Some models and advanced features can still have separate allowances. |
Luna and Sol Are Not the Same Upgrade
The model names make this sound simpler than it is. Free and Go users get GPT-5.6 Luna. OpenAI describes Luna as the fastest and lowest-cost model in the GPT-5.6 family. The new Think button also uses Luna. It does not quietly unlock GPT-5.6 Sol.
Sol is the stronger paid model for more difficult work. OpenAI says it is designed for complex coding, research, knowledge work and deeper reasoning. Plus and Pro users also get a slider that controls how much reasoning effort ChatGPT uses on an answer.
That creates a cleaner consumer split than a list of model names suggests. Free users get a much less restrictive everyday assistant. Paid users are still paying for stronger reasoning and broader tool access, not simply for the right to keep typing.
The Think Button Is the Bigger Interface Change
The Think button looks small. It may be the smartest part of the update.
Most people do not want to learn which model is good at which task. They know whether a question feels easy or hard. “Write this email” is easy. “Compare these three contracts and tell me where the risk sits” is harder. A button that says Think maps to that human distinction better than a model picker full of names.
OpenAI is making the same move on paid plans with the reasoning slider. Keep it low when you want speed. Turn it up when you want the model to work through more steps. The technology underneath is becoming more complicated while the interface is trying to become simpler. That is the real product story here.
Do Not Turn 68% Into “68% More Accurate”
OpenAI also published a strong accuracy number, and this is where headlines can go wrong. In an internal evaluation covering financial, medical and legal prompts that required factual detail, OpenAI says responses containing at least one factual error were about 62% less common with Luna and 68% less common with Sol than with GPT-5.5 Instant.
That does not mean Sol is “68% more accurate.” It does not mean it gets 68% more questions right. And it certainly does not mean finance, medical or legal answers are now safe to accept without checking. The metric is narrower: the share of evaluated responses containing at least one factual error fell by that amount in OpenAI’s own test.
Still, this is the kind of improvement that matters in daily use. Dates, numbers, rules and source-based questions are exactly where a confident wrong answer can cost you time or money. The right takeaway is fewer factual mistakes in OpenAI’s evaluation, not the end of factual mistakes.
OpenAI’s Own Documentation Still Has One Loose Edge
There is one detail we would watch before describing the paid experience too neatly. OpenAI’s launch post says the same GPT-5.6 Sol model now powers both Instant responses and deeper reasoning for Plus and Pro users.
But the Help Center page updated around the launch still labels the Instant option as GPT-5.5 Instant, with Medium, High and Extra High powered by GPT-5.6 Sol. It also says ChatGPT can automatically switch from Instant to Medium for complex requests.
Those descriptions are not fully aligned. That may get cleaned up quickly, but for now we would not tell a paid user that every quick reply is definitely Sol without qualification. The safe consumer reading is that Sol powers the paid reasoning experience and can be used automatically or directly, while the visible Instant label is still described differently in OpenAI’s support documentation.
What This Means for ChatGPT Go
Go is where the business model gets interesting. It is the lower-cost paid tier and is available in ChatGPT-supported countries, including the UAE. Before this announcement, extra message access was one of the obvious reasons to move up from Free. Unlimited text makes that reason much less important.
Go still has a job. OpenAI currently describes it as offering expanded image generation, more file uploads, more advanced data analysis, longer memory, and access to projects, tasks, custom GPTs and Library where available. If you use those features heavily, Free and Go are still different products.
But if your use is 90% typing, the calculation changes. That is good news for consumers and mild pressure on OpenAI to make every paid tier earn its price through capability rather than scarcity.
Why This Matters in the UAE
OpenAI explicitly lists the UAE as a supported location for GPT-5.6 when the model is available for the user’s plan and product. That makes this a real local rollout, not a US feature we are waiting to see someday.
For students, office workers, freelancers and small businesses, unlimited text removes friction from the most common workflows: drafting, translating, brainstorming, research follow-ups, planning and document preparation before a file ever needs to be uploaded.
It does not remove the need to check local facts. If you use AI for Arabic, Hijri dates, UAE addresses or local context, our real-world Arabic AI test is still the useful counterweight to any model-launch claim. A better model can reduce mistakes without making local verification obsolete.
The competitive pressure is also useful. We recently looked at Google’s latest push against ChatGPT. When one major assistant removes a core usage limit, rivals have to answer with better free access, better tools or a clearer reason to pay. UAE users benefit from that fight.
ChatGPT Is Becoming a Place, Not Just a Model
There is another reason unlimited text matters. ChatGPT is no longer only a box that answers questions. It is becoming a place where other services show up.
We have already seen that with Bayut inside ChatGPT and with the wider move toward AI-assisted shopping and payments. The more things happen inside the same conversation, the more annoying a basic text ceiling becomes. Removing it makes the chat itself a stronger home screen for everything OpenAI wants to add later.
That is why this update matters more than “free users get more messages.” OpenAI is removing friction from the layer that connects all of its other features.
Should You Still Pay?
If you only pay because you hate running out of text messages, wait and test the new Free experience when it reaches your account. You may find that the problem you were paying to solve has disappeared.
If you regularly upload files, generate images, run deeper analysis, need stronger reasoning, use longer workflows or depend on paid-only features, this announcement does not erase the value of Plus or Pro. It simply makes the dividing line clearer.
Free is becoming generous for conversation. Paid is increasingly about capability. That is a healthier split than making ordinary users count messages.
The Bottom Line
OpenAI has made one of the least flashy changes to ChatGPT and one of the most useful. Free users are getting a newer default model and, starting next week, the ability to keep ordinary text conversations going without the old rate limit.
The fine print is straightforward. Unlimited means text. Think still uses Luna. Tools still have limits. Sol remains part of the paid reasoning experience. And OpenAI’s 68% figure is an internal factual-error reduction measure, not a universal accuracy score.
The larger shift is easier to miss. OpenAI is slowly hiding the machinery. Free users get Chat and Think. Paid users get a reasoning slider. The model names matter less. What you are trying to do matters more. For most people, that is exactly how AI should work.
Sources
- OpenAI: August 6 product announcement covering GPT-5.6 Luna for Free and Go, unlimited text chats, Think, the Sol update and the internal factual-error evaluation. https://openai.com/index/improving-gpt-5-6-sol-in-chatgpt/
- OpenAI Help Center: GPT-5.6 in ChatGPT, including plan availability, Luna vs Sol, Think behavior, reasoning options, UAE eligibility and usage-limit notes. https://help.openai.com/en/articles/20001354-gpt-56-in-chatgpt
- OpenAI Help Center: ChatGPT Go features and global availability. https://help.openai.com/en/articles/11989085-what-is-chatgpt-go
- OpenAI Help Center: ChatGPT supported countries, confirming United Arab Emirates availability. https://help.openai.com/en/articles/7947663-chatgpt-supported-countries
- TechCrunch: Independent reporting on the Free and Go unlimited-text rollout, Think button and remaining tool limits. https://techcrunch.com/2026/08/06/openai-brings-unlimited-chatgpt-text-chats-to-free-users/
- The Verge: Independent reporting on the rollout timing, Luna default and Sol changes for paid users. https://www.theverge.com/ai-artificial-intelligence/976239/openai-chatgpt-free-go-text-chats
Robius.news — Dubai, UAE — 2026 | Built to be first. Built to be trusted.
