If you use the free version of ChatGPT and quietly assumed the paid version is where the real work happens, this is the week that assumption got weaker. On July 30, 2026, OpenAI cut the price of its GPT-5.6 Luna model by up to 80 percent. A few days later, it announced that ChatGPT had crossed one billion weekly active users and more than two million businesses. In the same window, free users started getting unlimited text chats on GPT-5.6 Luna along with a new "Think" button. None of this needs you to touch a line of code. It changes what your free ChatGPT account can do for you starting this week.
What actually happened, in order
Three things landed close together, and it helps to separate them.
First, the price cut. OpenAI dropped GPT-5.6 Luna from $1 to $0.20 per million input tokens, and from $6 to $1.20 per million output tokens, roughly an 80 percent reduction. Its larger model, GPT-5.6 Terra, got a smaller 20 percent cut. OpenAI said the savings came from efficiency work on its own serving systems, not a change in what the model can do. This is a developer-facing number, but it matters indirectly: a model this much cheaper to run shows up faster and cheaper in the apps and tools you already use, not just in ChatGPT itself.
Second, the scale milestone. OpenAI confirmed it now has over one billion active users and more than two million businesses on its models, up from 900 million weekly users reported back in February. That is a large jump in six months, and it tells you AI use at work has stopped being an early-adopter habit. A large share of the people you work with are now using it too, whether they mention it or not.
Third, and most useful for your actual workday: starting the first week of August, OpenAI made GPT-5.6 Luna the default model for free ChatGPT users, uncapped text chats that used to hit a daily limit, and added a "Think" button that gives the model more time to reason on a harder question.
The free tier of a major AI tool getting a real upgrade is worth more attention than another flagship model launch. Flagship launches change what is possible. Free tier upgrades change what millions of ordinary users, including the people on your team who never signed up for a paid plan, can actually do tomorrow morning.
What changed for someone on the free plan
If you or your team use the free version of ChatGPT, here is what is different starting this week.
Unlimited text chats
The daily cap on text conversations is gone for free users. You can now go back and forth on a drafting task, a summary, or a back-and-forth planning conversation without hitting a wall midway. File uploads, image generation, and some other tools still have limits, so this applies specifically to plain text conversation.
A "Think" button for harder questions
Free users can now tap Think on a question that needs real reasoning, a tricky comparison, a multi-step calculation, an ambiguous email you are not sure how to respond to, and the model will take longer to work through it instead of firing back the fastest plausible answer. Paid plans still get a more continuous depth control, but this is the first time free users get any version of that lever at all.
Fewer factual errors, by OpenAI's own account
OpenAI said internal evaluations on financial, legal, and medical questions found responses with at least one factual error were around 62 percent less common on the updated model compared with its predecessor. That is OpenAI's own internal testing, not an independent audit, so treat it as a reported improvement, not a guarantee. The habit that matters here has not changed: verify anything with a number, a date, or a legal or medical claim before you use it.
Why this is worth five minutes of your week
Three practical reasons this matters even if you never look at pricing pages.
One, if you or your team have been avoiding the free plan because it felt limited, that reason just got weaker. A finance executive, an HR manager, or an operations lead on the free tier now hits fewer of the daily walls that used to push people toward a paid subscription out of frustration rather than genuine need.
Two, cheaper models tend to spread. When a company cuts its own serving costs by 80 percent, that saving eventually shows up in the tools built on top of it: your CRM's AI summary feature, your email client's draft assistant, the chatbot on a vendor's website. Expect more of the software you already use to quietly get an AI feature added at no extra charge in the coming months.
Three, scale changes expectations. When over a billion people use a tool weekly, "I do not really use AI" stops being a neutral position at work. Colleagues and managers will notice the gap between people who have built a habit with it and people who have not.
What this does not mean
It is worth being precise here. A cheaper, faster, less error-prone model is still a model that can be confidently wrong, especially on anything specific: a case citation, a tax provision, a figure from a report it has not actually seen. The 62 percent error reduction is a real, reported improvement, not proof that verification is now optional. If anything, a model that sounds more assured and answers faster deserves slightly more scrutiny from you, not less, because a wrong answer delivered with total confidence is the version that is easiest to trust by mistake.
What to actually do this week
Three small, concrete steps. One, if you are on the free ChatGPT plan, open it and check whether the Think button and unlimited chats have reached your account yet, OpenAI is rolling this out gradually. Two, try the Think prompt above on one genuinely hard question from your actual work this week and compare the result to what you would normally get. Three, if your organisation runs on the free tier for casual use, mention this update in your next team huddle. Most people will not have noticed it, and a five-minute nudge is often the difference between a tool that sits idle and one that becomes a daily habit.
Knowing which product update is worth five minutes of your attention and which is noise is itself a skill, and it is one part of what we measure in the AiM AI Fluency Test: not how many AI tools you have heard of, but how well you actually use the ones already sitting on your phone and laptop.
The bigger pattern
This is the third update in as many weeks that changes what "using AI at work" costs or covers, after moves on office add-ins and content labelling rules. None of these is dramatic on its own. Together, they point one way: the gap between the free version of a major AI tool and the paid one is narrowing, and the tools keep getting cheaper to run. Professionals who notice these small shifts and adjust their habits within days, not months, benefit the most. That habit of staying current and actually changing how you work is exactly what we build in the AiM AI fluency courses.