OpenAI is tearing down one of the last major restrictions on its most popular product. Starting next week, the company is rolling out unlimited ChatGPT text chats for every user, including those on the free tier, ending the rate limits that have long pushed casual users toward a paid subscription. The change lands just as ChatGPT, according to OpenAI, recently crossed the 1 billion weekly user mark, making this one of the broadest usability upgrades the company has made in months.
Summary
Key takeaways
- OpenAI is removing limits on text-based chats for all ChatGPT users, including the free tier, with the change rolling out next week.
- GPT-5.6 Luna becomes the default model for Free and Go users this week, replacing GPT-5.5.
- A new “Think” button lets Free and Go users unlock higher reasoning power for harder questions, also arriving next week.
- Plus and Pro subscribers get an upgraded GPT-5.6 Sol model, available today, plus a thinking slider to control response depth.
- OpenAI says factual errors dropped 62% for GPT-5.6 Luna and 68% for GPT-5.6 Sol compared with GPT-5.5-Instant.
OpenAI removes text-chat limits for all ChatGPT users
OpenAI is scrapping rate limits on text conversations across every subscription tier, a shift that directly targets the friction free users have faced for years. Until now, anyone chatting heavily with ChatGPT on the Free or Go plans could hit a wall mid-conversation and get nudged toward a paid upgrade. That barrier is coming down for plain text exchanges specifically.
Unlimited text chats introduced for Free and Go users
The unlimited-chat policy, along with a new reasoning tool, is set to arrive next week for Free and Go subscribers. It’s a meaningful concession from OpenAI, since text-based conversation is the core use case for the vast majority of ChatGPT’s user base, and removing throttling there could keep more casual users engaged without ever touching a paid plan.
Limits remain on files, images, voice, and image generation
The unlimited policy has a clear boundary. OpenAI confirmed that separate caps will continue to apply to file uploads, image attachments, voice interactions, and image generation. In other words, the door opens wide for typing and reading, but everything involving richer media still runs into the same kind of throttling users have dealt with before.
New GPT-5.6 Luna model and Think button enhance experience for Free and Go users
Free and Go users are getting a new default engine this week: GPT-5.6 Luna, which replaces GPT-5.5 as the model powering everyday conversations on those tiers. It’s the backbone behind the newly unrestricted chat experience and represents a direct upgrade for the tiers that previously ran on older, more limited models.
GPT-5.6 Luna replaces GPT-5.5 as default for Free and Go tiers
Swapping in GPT-5.6 Luna as the standard model matters because it changes the baseline experience for hundreds of millions of people who never pay for ChatGPT. Rather than reserving newer models exclusively for paying customers, OpenAI is pushing its latest architecture down to the free tier, a strategic move that could widen the gap between ChatGPT and rival chatbots that keep stronger models locked behind paywalls.
Think button enables higher reasoning power on complex questions
Alongside the model swap, OpenAI is introducing a Think button for Free and Go users, letting them opt into higher reasoning power when a question demands more than a quick answer. According to OpenAI, the feature lets users “access higher reasoning for harder questions,” giving free-tier users a lightweight way to trigger deeper analysis without needing a subscription. That feature is scheduled to go live next week alongside the unlimited chat rollout.
Upgrades for Plus and Pro users with GPT-5.6 Sol and thinking slider
Paying subscribers aren’t left out. Plus and Pro users are getting an upgraded GPT-5.6 Sol model built for quick tasks such as answering questions, doing web research, offering advice, planning, writing, and helping with decisions. OpenAI says the updated model produces more compact, robust answers with tighter formatting, avoiding “extra detail when it does not help.”
GPT-5.6 Sol delivers faster, more compact, and robust responses
OpenAI describes the refreshed Sol model as designed to make fewer mistakes, “especially when answers depend on dates, numbers, sources, rules, or assumptions,” by making better use of the sources it retrieves when answering a question. Notably, this update applies to the general ChatGPT version of GPT-5.6 Sol only; the separate version of Sol used for Codex and Work remains unchanged.
Thinking slider and availability for Plus and Pro users
Plus and Pro subscribers also get a new slider that lets them decide how much thought ChatGPT puts into a given answer, tuning it up or down depending on how complex or multi-step the query is. Both the upgraded GPT-5.6 Sol and the thinking slider are available starting today for Plus and Pro users, making this the first piece of the update to actually ship.
Improved accuracy and performance across GPT-5.6 models
OpenAI is backing these releases with accuracy numbers pulled from its own internal evaluation. Compared with GPT-5.5-Instant, the company reports that factual errors were 62% less common for GPT-5.6 Luna and 68% less common for GPT-5.6 Sol. Those figures suggest a substantial jump in reliability for models now sitting at the center of both the free and paid ChatGPT experience.
Why this matters: accuracy has been one of the most persistent criticisms of AI chatbots, and if OpenAI’s internal numbers hold up in real-world use, pairing unlimited access with fewer factual slip-ups could meaningfully change how much people trust ChatGPT for research, planning, and everyday questions rather than treating it as a tool that needs constant double-checking.
ChatGPT user growth milestone
The timing of this rollout lines up with a notable growth marker: ChatGPT has recently surpassed 1 billion weekly active users, according to OpenAI. Removing text-chat limits right after clearing that threshold signals a push to keep that expanding user base engaged rather than bottlenecked by rate limits, especially as competition in the AI chatbot space keeps intensifying.
For an industry watching how far AI companies will go to widen free access without cannibalizing paid subscriptions, this update offers a clear signal: OpenAI is betting that unlimited conversation, paired with a smarter default model, is worth more than the incremental revenue it might have squeezed from rate-limited free users.
FAQ
What changes has OpenAI made for free ChatGPT users?
OpenAI removed text-chat limits and introduced the GPT-5.6 Luna model as default, along with a new Think button for complex questions.
What new features do Plus and Pro ChatGPT users get?
Plus and Pro users get an upgraded GPT-5.6 Sol model with faster and more compact responses, plus a thinking slider to control response complexity.
Are there still any limits on ChatGPT usage after the update?
Yes, limits remain for files, images, voice, and image generation, but text chats are now unlimited.
How much more accurate are the new GPT-5.6 models compared to previous versions?
GPT-5.6 Luna has 62% fewer factual errors, and GPT-5.6 Sol has 68% fewer errors compared to GPT-5.5-Instant.
Article produced with the assistance of artificial intelligence and reviewed by the editorial team.
