Text chat limits are gone — here’s what changes for your team

You know the rhythm. It’s late afternoon, a customer sends a long, rambling complaint. You open ChatGPT, paste it in, and ask for a calm, professional reply in Bahasa Malaysia and English. The draft is decent. Then the conversation stops responding. Your quota ran out. The reply sits in a draft tab until tomorrow morning, because you still have invoices to chase and a supplier to call.

That invisible ceiling — the message limit on free ChatGPT — has quietly shaped how many Malaysian SMEs treat AI. Not as a working tool, but as something you save for a rainy day. Your staff see it as a novelty. The result is a gap between what AI can actually do and what you manage to get out of it.

That gap just got a lot smaller.

The short version: OpenAI has removed the text chat limit for all ChatGPT users. Free and Go users now get GPT-5.6 Luna as the default model, plus a new “Think” button for complex questions. Plus and Pro users get an upgraded GPT-5.6 Sol model. Internal evaluations show factual errors dropped 62% and 68% on those two models respectively. Limits on files, images, and voice remain — but plain conversation is now open-ended. The changes arrive this week, with unlimited text chats rolling out next week.

What this actually means

For the past two years, free ChatGPT users lived inside a rationing mindset. Every message felt numbered. So you saved your conversations the way you save phone battery — you avoid opening the tool unless you really need it. That behaviour cost more than you realised. Small businesses learn by asking lots of questions, and the quota was quietly discouraging that.

Unlimited text chats remove the meter. You can now hold a long, multi-hour conversation about one project without watching a counter. The experience becomes closer to having a colleague who stays in the room with you until the work is done.

The new models also attack a different problem: reliability. In OpenAI’s evaluation, factual errors were 62% less common on GPT-5.6 Luna and 68% less common on GPT-5.6 Sol compared to the previous GPT-5.5-Instant. That matters when a confident but wrong answer could send you chasing a non-existent requirement or quoting a client incorrect delivery terms.

What this means for Malaysian SMEs specifically

Malaysian SMEs run on languages. You’re replying to a customer in Bahasa Melayu, negotiating with a supplier in English, reading a partner’s message in Chinese, and keeping the accountant copied in a mix of all three. Until now, the chat limit forced you to choose: use the conversation for the customer, or save it for your own proposal? That choice no longer exists. You can keep separate windows open for each language, each project, each relationship — and none of them will cut you off mid-thought.

Think about a realistic day. You run a food supplier with eight staff. Your WhatsApp Business line is full of order enquiries. Your Shopee and Lazada listings need sharper product descriptions. A new client wants a proposal by Friday. A vendor changed their payment terms and you need to compare the old and new clauses. None of those are quick tasks. All of them are text-based conversations. With the cap removed, you can work through each one in a dedicated thread — and the model remembers the context as the thread grows. That continuity is where the actual time saving lives.

The Think button deserves more attention than it’s getting. It tells the model to reason harder for complex questions — the kind where a fast answer isn’t good enough. Planning a workflow change, evaluating whether to expand your delivery zones, breaking down the structure of a government grant application: these are deliberate, multi-step tasks. Previously, that deeper reasoning mode was reserved for paying users. Now it’s available to everyone. It’s a “slow down and be careful” switch for decisions where a sloppy answer costs you real hours.

There’s also a quieter consequence: accuracy compounds. When your operations lead summarises a safety guideline, or your admin staff drafts a response about a delayed shipment, small errors stack up across the week. The measured drop in factual errors directly cuts the time your team spends double-checking the AI’s work — and the social cost of sending a client confidently incorrect information.

The text limit was never just a technical ceiling. It was a signal that AI was a side tool — something you dip into now and then. Removing it changes that signal: you’re now allowed to build a working routine around it.

A quick snapshot of what changed

What changed	Who gets it	Why it matters to you
Unlimited text chats	All users	No more rationing; long working conversations now possible
GPT-5.6 Luna as default model	Free and Go users	62% fewer factual errors vs GPT-5.5-Instant
New “Think” button	Free and Go users	Deeper reasoning for multi-step questions
GPT-5.6 Sol update	Plus and Pro users	68% fewer factual errors; faster on research and planning
Thinking slider	Plus and Pro users	You choose how much effort the model applies
File, image, voice limits	Still capped	Text is unlimited; media uploads aren’t

Source: TechCrunch, 6 August 2026

Practical takeaways — what to do this week

• Retire the rationing habit. If you’ve been saving ChatGPT for “important” tasks, stop. The quota is gone. Waste a conversation on a rough idea — that’s how you’ll start using it properly.
• Open one thread per project. One for customer replies, one for proposals, one for planning. Long conversations now work in your favour; the model remembers what was said.
• Use the Think button for decisions, not drafts. Reserve it for process changes or business choices where a lazy answer wastes real time.
• Tell your team the limit is gone. Most staff still behave like the old cap exists. A quick message in your group chat changes their default behaviour.
• Plan around media caps. Files, images, and voice remain limited. If your workflow depends on uploading many documents, spread those uploads out rather than assuming everything is now unlimited.

The bigger picture

The direction of travel is clear. ChatGPT recently crossed 1 billion weekly users, and OpenAI’s moves — removing limits, improving accuracy, widening access to deeper reasoning — point toward a future where the free tier is the front door, not a trial offer. For AI providers, scale is the strategy. For users, the tool becomes a quieter, more reliable presence in the working day.

For Malaysian SMEs, the implication is straightforward. The tools are no longer the bottleneck. Your workflows are. A business that builds a habit of drafting, reviewing, and planning with AI this quarter will be in a different position next year than one that still treats it as an occasional novelty.

So pick one process you keep postponing because it requires “writing” — the proposal, the standard operating procedure, the weekly supplier update — and hand it to ChatGPT this week. The conversation will no longer run out on you. Whether you make something of that is now entirely up to you.

Ready to Streamline Your Operations?

Your business should run itself. AutoRunBiz deploys AI agents to automate your daily operations — WhatsApp orders, invoicing, customer follow-ups, and accounting. Book a free 15-min ops audit to see where automation fits your business →