# ChatGPT just dropped text limits for free users — here’s what really changed

By Rafiqul Islam Rabbi · AI · Published Fri, 07 Aug 2026 09:02:27 GMT · Updated Fri, 07 Aug 2026 15:02:27 GMT
Source: The Current Tribune — https://currenttribune.com/article/chatgpt-unlimited-text-chats-gpt-5-6-luna

ChatGPT just pulled the cap off its most basic feature. OpenAI is removing limits on text-based chats for free users and its mid-tier Go plan, making the core ChatGPT experience truly unlimited for the first time. The shift is powered by a new model, GPT-5.6 Luna, and it’s paired with a new Think button that lets you tap into extra reasoning power when you actually need it.

## Unlimited chats, but not unlimited everything

The headline change is simple: text chats are now uncapped for everyone on ChatGPT’s free and Go tiers. Previously, free users hit soft limits on how many prompts they could run in a given period before being throttled or nudged to upgrade. Those guardrails are going away for plain text conversations.

There are important boundaries, though. OpenAI is only lifting limits for text. Files, images, voice, and image generation still have their own separate caps, so you can’t suddenly run infinite audio transcriptions or spin up endless AI-generated pictures on a free account. This is about the core back-and-forth chat, not the entire feature set.

To deliver this at scale, OpenAI is rolling out a new default engine for these tiers: GPT-5.6 Luna. It replaces GPT-5.5 as the standard model that powers free and Go chats. You won’t have to pick it from a menu — it simply becomes the model you’re talking to whenever you open ChatGPT without a paid Plus or Pro subscription.

## GPT-5.6 Luna: OpenAI’s new everyday default

GPT-5.6 Luna is pitched as a direct upgrade over GPT-5.5, with a particular focus on accuracy. In OpenAI’s own internal testing, GPT-5.6 Luna produced 62% fewer factual errors than GPT-5.5-Instant, one of the prior fast models used across ChatGPT. Those numbers are lab results, not a guarantee, but they hint at why OpenAI is confident enough to open the floodgates on text usage.

For the average free user, this means two things. First, you’re chatting with a more capable model by default — one that should be better at staying on topic, getting basic facts right, and handling everyday planning, writing, and Q&A. Second, you can push it harder without worrying about burning through a quota of messages.

The move also has a bigger strategic signal: OpenAI is clearly leaning into ChatGPT as a mass-market utility. Crossing 1 billion weekly users and then dropping the main friction for new or casual users sends a message to rivals that usage, not stingy feature gating, is the priority.

## The new Think button for complex queries

Unlimited chats would already be a big deal, but OpenAI is also changing how you ask tougher questions. Free and Go users are getting a new Think button that switches the model into a higher-reasoning mode for complex prompts.

Instead of every request consuming maximum compute and time, you can tap Think when you’re dealing with something that genuinely needs extra steps — like multi-part research questions, involved planning, or tricky logic problems. In other words, the default is fast and efficient; Think is the manual override when you care more about depth than speed.

This matters because “reasoning” has become the buzzword around modern AI models. Users don’t just want plausible-sounding answers; they want the system to actually work through a problem, consider alternatives, and land on something that holds up. Giving free users a visible control for that, instead of burying it behind a paywall, is a notable shift.

### When you’ll actually use Think

The Think button is overkill for simple one-shot questions — definitions, quick lists, short emails, and the usual homework clean-up. Where it should shine is on prompts that naturally break into steps: planning a multi-city trip with budget constraints, designing a multi-week study schedule, or comparing several options based on criteria you care about.

It also gives power users a way to consciously trade a bit of latency for better reasoning, rather than rolling the dice on whatever default behavior the model chooses in the background.

![user exploring new Think button in ChatGPT unlimited text interface on laptop](/media/2026/08/chatgpt-unlimited-text-chats-gpt-5-6-luna-inline.webp)
*The new Think control lets ChatGPT users trade speed for deeper reasoning when it matters. (Photo: Ars Electronica / BY-NC-ND via Openverse)*

## Paid tiers get GPT-5.6 Sol — and a thinking slider

While free and Go users move to GPT-5.6 Luna, ChatGPT Plus and Pro subscribers are getting an upgraded GPT-5.6 Sol model that’s tuned for quick tasks like question answering, web research, planning, advice, writing, and decision help.

The new Sol flavor is designed to give more compact, robust answers — basically, fewer rambly paragraphs and more focused responses. In OpenAI’s internal evaluation, Sol cut factual errors by 68% compared to GPT-5.5-Instant, edging out Luna on accuracy.

On top of that, Plus and Pro users are getting a more granular version of the new reasoning control: a thinking slider. Instead of a single Think button, subscribers can adjust how much thought the model applies to a query, dialing it up for harder problems or down when they just need a quick, rough answer.

That slider reflects something longtime users have been asking for: more transparency and control over how the model spends its compute on their behalf. Rather than hoping the system auto-detects when a question is “hard,” you can tell it directly.

It’s worth noting that the Sol model being upgraded here is specifically the one used in the main ChatGPT experience; OpenAI clarifies that the version of GPT-5.6 Sol used for its Codex and Work offerings is unchanged.

## Rollout timing: when you’ll see the changes

If you’re on ChatGPT Plus or Pro, the upgraded GPT-5.6 Sol is available now. You should already see the new model option and the thinking slider in your account if the update has hit your region.

Free and Go users don’t have long to wait. OpenAI says the new default GPT-5.6 Luna model and the Think button are rolling out this week, with unlimited text chats arriving next week. Once it lands on your account, you’ll be able to fire off as many text prompts as you want without running into a hard cap.

The staggered rollout is typical for big ChatGPT updates and helps OpenAI manage the sudden spike in usage that unlimited text inevitably invites.

## Why OpenAI is uncapping ChatGPT now

Dropping message caps right after surpassing 1 billion weekly users might look backward at first — why loosen limits once growth is already sky-high? But it fits a broader pattern: OpenAI increasingly wants ChatGPT to be the default interface for everyday AI use, not a constrained demo funneling people into paid plans.

Unlimited text chats lower the psychological barrier to experimenting with more use cases. If you don’t have to ration prompts, you’re more likely to treat ChatGPT like a standing assistant for small, frequent tasks — rewriting lines of text, brainstorming ideas, checking logic — instead of a tool you only open for “important” questions.

At the same time, OpenAI isn’t giving away the whole store. Heavy-duty features like large file handling, voice, and image generation still sit behind distinct limits and, in practice, push power users toward paid tiers. The company can afford to be generous with plain text because that’s where efficiency gains from newer models like Luna are easiest to realize.

### Pressure on rivals

This update also raises the bar for competitors. Other AI chatbots have dangled generous free tiers, but many still impose tight caps on high-quality models or reserve advanced controls for paying users. By making its mainstream model unlimited for text and layering in a reasoning toggle, OpenAI is resetting expectations for what “free” should mean in this space.

If users get used to having no text limits and a visible knob for deeper thinking, rivals that cling to strict quotas or opaque behavior risk feeling dated fast.

## What This Means

For most people, this update makes ChatGPT less of a special-occasion tool and more of an everyday utility. Unlimited text chats with GPT-5.6 Luna remove the friction that made some users hold back, while the Think button gives you a simple, visible way to ask for more rigorous reasoning when you actually need it.

Paid users don’t get left behind, either. The upgraded GPT-5.6 Sol model and its thinking slider underline that Plus and Pro remain the place to go for faster, more controlled work — especially if you live in ChatGPT for research, planning, or writing.

The bigger takeaway is that OpenAI is betting hard on depth of engagement over strict metering. If the company can keep scaling models like Luna and Sol while holding error rates down, uncapping text chat may turn ChatGPT from an occasional curiosity into the default interface for how a billion people think out loud with an AI every week.

*Photo: Ars Electronica / BY-NC-ND via Openverse*