OpenAI Just Made ChatGPT Free and Unlimited for Everyone. If You Charge for 'I'll Set Up Your AI Assistant,' Your Pitch Just Got Harder.
On August 6, OpenAI announced that free ChatGPT accounts are getting unlimited text chats, starting the week of August 10. The old cap, 10 messages every 5 hours, is gone. No credit card, no waitlist, no catch beyond the model doing the answering. ChatGPT already has something like 1 billion weekly users. Removing the last hard limit on the free tier isn't a capability upgrade, it's a distribution move, and it lands directly on a specific slice of solo-operator income: the "I'll help you use ChatGPT" consulting gig.
I want to be precise about what changed and what didn't, because the headline invites overreaction in both directions.
What actually shipped
GPT-5.6 Luna becomes the default model for Free and Go tier users, replacing GPT-5.5, and those users get unlimited text conversations starting the week of August 10. Free and Go users are also getting a new "Think" button that lets them opt into higher reasoning effort for harder questions, arriving alongside the unlimited-chat rollout. Plus and Pro subscribers get a more granular version of the same idea, a thinking slider that lets them tune how much reasoning effort a query gets, plus an upgraded GPT-5.6 Sol model tuned for quick tasks like research, planning, and writing. What doesn't change for the free tier: file uploads, images, voice, and image generation stay under separate limits, and free chats still train future models by default, which was already true.
The framing matters here. This is OpenAI closing a gap, not opening a lead. Claude and Gemini have both offered unlimited free-tier text chat for a while now. A user comparing "unlimited free chat" across the three major labs previously had two good options and one capped one. Now there are three unlimited options, and the one with the largest existing user base just removed its own weakest point of comparison.
Why this is a distribution story, not a model story
Nobody is switching to ChatGPT because Luna is dramatically smarter than what came before. The move is about retention and defense: with three frontier labs now offering unlimited free text, the deciding factor for a huge chunk of casual users stops being "which one lets me chat as much as I want" and starts being "which one do I already have installed, already know how to use, or heard about from a friend." OpenAI is playing that game from a position of enormous scale advantage, and removing a usage cap costs them very little marginal compute for casual conversational use compared to the retention value of not losing users to a competitor's unlimited tier.
That's a rational move for OpenAI. It's also the kind of move that quietly erodes a specific business model that a lot of solo operators have been running without fully naming it: the "AI setup and onboarding" service.
The actual consequence for solo consultants
If part of your income has been showing non-technical clients (small business owners, freelancers, family members running a side business) how to use ChatGPT, that offer just lost a chunk of its perceived value. The client's objection used to sometimes be "but I'd have to pay for it to actually be useful" or "the free version stops working after a few questions." That objection is gone. A client can now sit down, open a free account, and chat with a genuinely capable model for as long as they want, with zero setup fee paid to anyone.
I've done versions of this work myself, and the honest read is that the part of the job that was "get them to unlimited access" was never the valuable part anyway. It just felt valuable because it was the visible, demoable part of the interaction. The actual expertise, knowing which prompts get useful output, understanding the client's workflow well enough to know what to automate, wiring the chat interface into their actual tools, was always the harder and more defensible skill. It just used to be bundled with "also I'll get you past the free tier limits," and that bundling made the whole package feel more substantial than it was.
Where the paid work still exists
The distinction that matters now is integration versus access. Access to a capable model is free, from three different vendors, with no meaningful catch for someone who just wants to chat. Integration, wiring an AI assistant into a client's actual calendar, email, CRM, or document workflow so it does useful work without them having to think about prompting, is still hard, still requires judgment about their specific business, and still isn't something a free unlimited chat tier delivers on its own.
If your current pitch to clients sounds anything like "let me get you set up with ChatGPT and show you how to use it," it's worth rewriting that pitch this week, before a client rewrites it for you by pointing out they don't need to pay for something that's now free. If your pitch already sounds like "let me build a workflow where AI reads your invoices and drafts your follow-up emails without you touching a chat window," nothing about this announcement threatens that, because that was never about unlimited chat access in the first place.
The honest counter-take
It's possible I'm overstating how much of the "AI onboarding" consulting market was ever priced around unlimited access rather than around hand-holding. A lot of non-technical clients don't actually hit usage caps in normal use, they're intimidated by the interface, unsure what to ask, or simply haven't tried it. For that segment, the value was never "I'll get you past the free tier," it was "I'll sit with you until you're comfortable," and that value doesn't disappear because the underlying tool got more generous. If your business is closer to hand-holding and confidence-building than to raw access-provisioning, this announcement barely touches you.
What I'd actually do
Look at your last five client conversations about AI setup and ask which ones were fundamentally about access versus fundamentally about integration or confidence-building. If most of them were access conversations, that revenue line is shrinking starting this week, and you have a decision to make about repositioning toward workflow work before the market makes it for you. If most of them were integration or hand-holding conversations, keep doing exactly what you're doing. The rollout starting August 10 doesn't touch either of those, only the narrow slice in between.
Author
Lukas
@lukcombinator