OpenAI says ChatGPT’s Free tier will receive unlimited text-only chats with GPT-5.6 Luna beginning next week, removing the ordinary prompt cap that has interrupted longer conversations and pushed users toward an upgrade screen. The change also applies to the $8-per-month ChatGPT Go tier, despite OpenAI framing much of its August 6 announcement around expanded access for Free users. The Verge and Engadget both reported the broader Free-and-Go rollout, while OpenAI’s own availability section makes that scope explicit.

For Windows users running the ChatGPT desktop app—or using ChatGPT in a browser—the practical effect is straightforward: ordinary back-and-forth text work such as drafting, troubleshooting, explaining PowerShell output, revising documents, or iterating on code should no longer consume a visible message allowance. But “unlimited” has a narrow definition here. Uploading a file, adding an image, generating images, and using other tools will still be subject to separate quotas, and OpenAI says the text access remains subject to abuse guardrails.

That distinction is the story. OpenAI has eliminated the limit on the cheapest and most frequent form of ChatGPT use, while retaining limits on the actions that cost more compute or involve more infrastructure. It has also made no announcement that advertisements, already being tested for Free and Go users in the United States, are going away.

A Windows desktop displays a GPT-5.6 Luna chat explaining quantum computing alongside subscription plan details.GPT-5.6 Luna Becomes the Free Default​

OpenAI is replacing GPT-5.5 Instant with GPT-5.6 Luna as the default model for ChatGPT Free and Go accounts this week. Luna is the fast, lower-cost member of the GPT-5.6 family, and OpenAI says it will provide unlimited text chats starting next week. Free and Go users will also receive a new Think button intended to give Luna more time to reason through harder prompts.

The model change matters as much as the quota change. Previously, a free user who reached the current text limit could be blocked from continuing with the main model until a reset time, while access to GPTs was tied to that same model allowance. OpenAI’s existing Free Tier FAQ still describes model usage as limited and says ChatGPT will notify users when they reach the current cap. That Help Center page was last updated before this week’s announcement, so it reflects the outgoing policy rather than evidence that OpenAI has reversed course.

The new policy does not mean every ChatGPT function has become limitless. OpenAI’s August 6 post specifically retains caps for file uploads, images, and other tools. Engadget adds that image creation and the recently updated voice experience remain separately constrained. A conversation that starts as ordinary text may therefore still encounter a limit once it becomes multimodal—such as attaching a Windows Event Viewer export, uploading a screenshot of an error dialog, or asking the service to generate an image.

OpenAI has not published a numerical replacement schedule for those remaining limits, nor has it provided a platform-by-platform rollout timetable for GPT-5.6 Luna on web, Windows, macOS, iOS, and Android. Its July desktop-app update confirms that ChatGPT and Codex experiences are available across Windows and macOS plans, but the August announcement only says Luna becomes the default for Free and Go users “this week.” Users should treat the in-app model behavior and account notices as the operational record until the rollout finishes.

The Word “Unlimited” Still Has a Guardrail Attached​

OpenAI’s announcement is unusually clear about the boundary but less specific about enforcement. The company says unlimited text chats are “subject to abuse guardrails,” without defining the thresholds, patterns, or remediation process. That leaves the usual caveat attached to virtually every consumer AI product advertised as unlimited: there may be no ordinary hourly prompt counter, but automated anti-abuse systems can still intervene.

For normal interactive use, that wording should not change much. It is aimed at sustained automated, scripted, spam-like, account-sharing, or otherwise abusive behavior—not a person asking for twenty rounds of edits to a deployment script. Still, it means OpenAI has not promised infinite throughput, no service throttling, or unrestricted use of the model for machine-driven workloads.

The policy also does not turn ChatGPT into an API substitute. ChatGPT plans and API billing are separate products with separate controls, and OpenAI has not announced unlimited GPT-5.6 API access. IT teams should not interpret this as permission to shift batch summarization, log parsing, or application integrations from API workflows into consumer chat sessions. Aside from governance and privacy concerns, that would run directly into product terms and the undefined abuse safeguards.

The more immediate benefit is to individual users whose work is conversational by nature: troubleshooting a sequence of Windows failures, refining a script across multiple iterations, walking through a configuration problem, or developing documentation without preserving messages for later. The removal of the prompt counter will make those sessions feel more continuous, even though they can still be interrupted by a tool-specific limit.

ChatGPT Go Loses Its Cleanest Selling Point​

The Free-versus-Go wording deserves scrutiny because OpenAI’s public materials are not equally precise. The headline and early sections of the August 6 post say OpenAI is expanding access for Free users; the formal availability section later says GPT-5.6 Luna will be the default for Free and Go accounts and that both will receive unlimited text chats next week. The Verge and Engadget independently reported the same Free-and-Go scope.

That leaves little doubt about eligibility, but it changes the economics of ChatGPT Go. When OpenAI launched Go globally in January, it positioned the $8 monthly plan around more messages, more file uploads, and more image creation than the Free tier. Its launch material cited ten times more messages, uploads, and images than Free, while advertising would still be present on Go.

Once Free accounts receive unlimited plain-text conversations with the same default GPT-5.6 Luna model, message volume is no longer the simple reason to pay for Go. Go can still be valuable for users who hit tool limits, need more uploads or image generation, or want expanded access beyond text prompts. But OpenAI has not stated what happens to the old “ten times more messages” comparison now that plain text itself becomes unlimited on both plans.

That missing detail is more than a marketing footnote. Existing Go subscribers are paying for a tier that also may show advertisements, and OpenAI has not said whether it will alter pricing, benefits, tool quotas, or ad treatment following this change. The company’s current ad documentation says ads may appear on Free and Go plans, while Plus, Pro, Business, Enterprise, and Education accounts do not receive them.

In other words, Go becomes primarily a multimodal-capacity plan rather than a text-chat plan. That may be a reasonable product split, but OpenAI has not yet laid it out in a revised plan comparison.


Ads Remain the Likely Trade for Free Access​

OpenAI introduced ChatGPT advertising tests for Free and Go users in the United States on February 9, 2026. Its earlier explanation of the program linked ads directly to broader access and fewer usage limits, saying advertising could help support a faster, more capable service without requiring every user to pay. The August 6 announcement now delivers the largest of those reduced-limit promises for text prompts.

OpenAI says ads are placed below ChatGPT responses and should not affect the model’s answer. Its advertising FAQ says ads can be informed by the current conversation and, when personalization is enabled, certain signals from a user’s wider ChatGPT experience. Advertisers are not supposed to receive users’ chat histories, memories, or personal information.

The unlimited-text announcement does not mention advertising once. It does not say whether the rollout will expand the ad test beyond its existing territories, whether more ads will appear as text use rises, or whether Free users who select an ads-free, lower-limit option will retain unlimited text chats. That last point is particularly important because OpenAI’s ads-free control for Free accounts has been described as a trade: no ads in exchange for reduced message limits and reduced feature access.

Until OpenAI updates that policy, users should not assume that “unlimited text chats” overrides every other account-level choice. The company’s own documentation now needs to reconcile an ads-free option built around fewer messages with a Free tier marketed as having no text-chat limit.

Plus and Pro Receive a Different Upgrade​

While Free and Go accounts gain access and volume, Plus and Pro accounts receive an update to GPT-5.6 Sol, OpenAI’s more capable ChatGPT model. OpenAI says Sol should provide more focused answers, use less unnecessary formatting, and make fewer factual mistakes where dates, numerical claims, sources, rules, and assumptions matter.

The company’s internal evaluation claims responses with at least one factual error were 62 percent less common with GPT-5.6 Luna and 68 percent less common with GPT-5.6 Sol than with GPT-5.5 Instant on financial, medical, and legal prompts requiring factual detail. Those are OpenAI’s own test results, not an independently replicated benchmark, and they should be read as a product claim rather than a guarantee that ChatGPT is safe for high-stakes decisions.

Plus and Pro subscribers also gain a slider that controls how much “thought” ChatGPT applies to an answer, available on web, mobile, and desktop. OpenAI says the updated Sol and slider are available now, while the Free and Go unlimited-text rollout is scheduled for next week.

The immediate takeaway for Windows users is simple: use Free ChatGPT for ongoing text conversations without treating it as an unlimited image, file-analysis, voice, or automation service. For Go subscribers, check actual upload and image usage before renewing—the plan’s old message-limit advantage is about to disappear, while its ads and tool quotas remain.


References​

  1. Primary source: The Verge
    Published: August 6, 2026 at 5:00 PM UTC
  2. Independent coverage: Engadget
    Published: August 6, 2026 at 5:01 PM UTC
  3. Related coverage: axios.com
  4. Related coverage: engadget.com
  5. Related coverage: help-lb.openai.com
  6. Related coverage: help.openai.com
  7. Related coverage: axios.com
  8. Related coverage: help.openai.com
  9. Related coverage: searchenginejournal.com
  10. Related coverage: searchengineland.com
  11. Related coverage: macrumors.com
  12. Related coverage: theregister.com
  13. Related coverage: sfchronicle.com
  14. Related coverage: deploymentsafety.openai.com
  15. Related coverage: cdn.openai.com
  16. Related coverage: cdn.openai.com
  17. Related coverage: tomsguide.com
  18. Related coverage: androidcentral.com
  19. Related coverage: techradar.com
  20. Related coverage: about.chat
  21. Related coverage: openai.com
  22. Related coverage: humai.blog
  23. Related coverage: techcrunch.com
  24. Related coverage: thesummerai.com
  25. Related coverage: adthena.com
  26. Related coverage: mavgpt.ai
  27. Related coverage: intuitionlabs.ai
  28. Related coverage: techxplore.com
  29. Related coverage: openai.com
  30. Related coverage: zapier.com