OpenAI is tearing down one of many final main restrictions on its hottest product. Beginning subsequent week, the corporate is rolling out limitless ChatGPT textual content chats for each person, together with these on the free tier, ending the speed limits which have lengthy pushed informal customers towards a paid subscription. The change lands simply as ChatGPT, in response to OpenAI, lately crossed the 1 billion weekly person mark, making this one of many broadest usability upgrades the corporate has made in months.
Key takeaways
- OpenAI is eradicating limits on text-based chats for all ChatGPT customers, together with the free tier, with the change rolling out subsequent week.
- GPT-5.6 Luna turns into the default mannequin for Free and Go customers this week, changing GPT-5.5.
- A brand new “Suppose” button lets Free and Go customers unlock larger reasoning energy for more durable questions, additionally arriving subsequent week.
- Plus and Professional subscribers get an upgraded GPT-5.6 Sol mannequin, obtainable as we speak, plus a considering slider to regulate response depth.
- OpenAI says factual errors dropped 62% for GPT-5.6 Luna and 68% for GPT-5.6 Sol in contrast with GPT-5.5-Instantaneous.
OpenAI removes text-chat limits for all ChatGPT customers
OpenAI is scrapping price limits on textual content conversations throughout each subscription tier, a shift that immediately targets the friction free customers have confronted for years. Till now, anybody chatting closely with ChatGPT on the Free or Go plans may hit a wall mid-conversation and get nudged towards a paid improve. That barrier is coming down for plain textual content exchanges particularly.
Limitless textual content chats launched for Free and Go customers
The unlimited-chat coverage, together with a brand new reasoning instrument, is ready to reach subsequent week for Free and Go subscribers. It’s a significant concession from OpenAI, since text-based dialog is the core use case for the overwhelming majority of ChatGPT’s person base, and eradicating throttling there may hold extra informal customers engaged with out ever touching a paid plan.
Limits stay on information, photographs, voice, and picture technology
The limitless coverage has a transparent boundary. OpenAI confirmed that separate caps will proceed to use to file uploads, picture attachments, voice interactions, and picture technology. In different phrases, the door opens huge for typing and studying, however all the pieces involving richer media nonetheless runs into the identical sort of throttling customers have handled earlier than.
New GPT-5.6 Luna mannequin and Suppose button improve expertise for Free and Go customers
Free and Go customers are getting a brand new default engine this week: GPT-5.6 Luna, which replaces GPT-5.5 because the mannequin powering on a regular basis conversations on these tiers. It’s the spine behind the newly unrestricted chat expertise and represents a direct improve for the tiers that beforehand ran on older, extra restricted fashions.
GPT-5.6 Luna replaces GPT-5.5 as default for Free and Go tiers
Swapping in GPT-5.6 Luna as the usual mannequin issues as a result of it modifications the baseline expertise for a whole bunch of tens of millions of people that by no means pay for ChatGPT. Slightly than reserving newer fashions solely for paying prospects, OpenAI is pushing its newest structure all the way down to the free tier, a strategic transfer that would widen the hole between ChatGPT and rival chatbots that hold stronger fashions locked behind paywalls.
Suppose button allows larger reasoning energy on advanced questions
Alongside the mannequin swap, OpenAI is introducing a Suppose button for Free and Go customers, letting them decide into larger reasoning energy when a query calls for greater than a fast reply. Based on OpenAI, the function lets customers “entry larger reasoning for more durable questions,” giving free-tier customers a light-weight option to set off deeper evaluation without having a subscription. That function is scheduled to go reside subsequent week alongside the limitless chat rollout.
Upgrades for Plus and Professional customers with GPT-5.6 Sol and considering slider
Paying subscribers aren’t omitted. Plus and Professional customers are getting an upgraded GPT-5.6 Sol mannequin constructed for fast duties corresponding to answering questions, doing net analysis, providing recommendation, planning, writing, and serving to with selections. OpenAI says the up to date mannequin produces extra compact, strong solutions with tighter formatting, avoiding “further element when it doesn’t assist.”
GPT-5.6 Sol delivers sooner, extra compact, and strong responses
OpenAI describes the refreshed Sol mannequin as designed to make fewer errors, “particularly when solutions rely on dates, numbers, sources, guidelines, or assumptions,” by making higher use of the sources it retrieves when answering a query. Notably, this replace applies to the overall ChatGPT model of GPT-5.6 Sol solely; the separate model of Sol used for Codex and Work stays unchanged.
Considering slider and availability for Plus and Professional customers
Plus and Professional subscribers additionally get a brand new slider that lets them resolve how a lot thought ChatGPT places right into a given reply, tuning it up or down relying on how advanced or multi-step the question is. Each the upgraded GPT-5.6 Sol and the considering slider can be found beginning as we speak for Plus and Professional customers, making this the primary piece of the replace to really ship.
Improved accuracy and efficiency throughout GPT-5.6 fashions
OpenAI is backing these releases with accuracy numbers pulled from its personal inside analysis. In contrast with GPT-5.5-Instantaneous, the corporate reviews that factual errors had been 62% much less frequent for GPT-5.6 Luna and 68% much less frequent for GPT-5.6 Sol. These figures counsel a considerable bounce in reliability for fashions now sitting on the heart of each the free and paid ChatGPT expertise.
Why this issues: accuracy has been one of the vital persistent criticisms of AI chatbots, and if OpenAI’s inside numbers maintain up in real-world use, pairing limitless entry with fewer factual slip-ups may meaningfully change how a lot folks belief ChatGPT for analysis, planning, and on a regular basis questions slightly than treating it as a instrument that wants fixed double-checking.
ChatGPT person progress milestone
The timing of this rollout traces up with a notable progress marker: ChatGPT has lately surpassed 1 billion weekly energetic customers, in response to OpenAI. Eradicating text-chat limits proper after clearing that threshold alerts a push to maintain that increasing person base engaged slightly than bottlenecked by price limits, particularly as competitors within the AI chatbot house retains intensifying.
For an trade watching how far AI corporations will go to widen free entry with out cannibalizing paid subscriptions, this replace provides a transparent sign: OpenAI is betting that limitless dialog, paired with a better default mannequin, is price greater than the incremental income it might need squeezed from rate-limited free customers.
FAQ
What modifications has OpenAI made totally free ChatGPT customers?
OpenAI eliminated text-chat limits and launched the GPT-5.6 Luna mannequin as default, together with a brand new Suppose button for advanced questions.
What new options do Plus and Professional ChatGPT customers get?
Plus and Professional customers get an upgraded GPT-5.6 Sol mannequin with sooner and extra compact responses, plus a considering slider to regulate response complexity.
Are there nonetheless any limits on ChatGPT utilization after the replace?
Sure, limits stay for information, photographs, voice, and picture technology, however textual content chats are actually limitless.
How far more correct are the brand new GPT-5.6 fashions in comparison with earlier variations?
GPT-5.6 Luna has 62% fewer factual errors, and GPT-5.6 Sol has 68% fewer errors in comparison with GPT-5.5-Instantaneous.
Article produced with the help of synthetic intelligence and reviewed by the editorial workforce.