On August 6, 2026, OpenAI announced a sweeping adjustment to its freemium structure, removing text chat rate limits for all free and Go-tier users of ChatGPT as the platform crosses 1 billion weekly active users. The update introduces GPT-5.6 Luna as the new default model for non-paying accounts, replacing GPT-5.5 and adding a “Think” button to trigger deeper reasoning on complex queries. This marks a decisive shift in how OpenAI manages capacity for its broadest user base: instead of relying on rate caps to restrict usage or drive paid conversions, the company is eliminating friction at the text layer while upgrading the underlying intelligence available to free accounts the announcement confirmed the removal of limits and the introduction of GPT-5.6 Luna as users surpass 1 billion weekly active users.
The technical rollout begins around August 10–13, 2026, when unlimited text volume reaches free and Go users alongside the Luna model deployment across non-paying tiers. Luna’s inclusion of a dedicated “Think” button gives the free experience access to chain-of-thought capabilities typically associated with premium reasoning loops. This change effectively bridges the performance gap for standard text interactions, allowing non-paying users to engage with a model that adapts its processing depth based on complexity without requiring a subscription. The removal of rate limits ensures these enhanced models can be used continuously, turning what was previously a capped luxury into an unrestricted utility the rollout of unlimited text volume and GPT-5.6 Luna will occur across Free and Go accounts starting around August 10–13.
For subscribers, the accompanying update focuses on precision rather than volume availability. Plus and Pro users gain access to GPT-5.6 Sol, a model variant OpenAI describes as specifically tuned for factual reliability. Paid tiers also receive a thinking slider that allows manual regulation of chain-of-thought processing depth per response, alongside tighter formatting controls for generated text. While free users shed rate limits for raw access and automated reasoning triggers, paying accounts receive granular controls over output structure and accuracy assurance, suggesting OpenAI’s value proposition for revenue-generating tiers is evolving toward customization and trust rather than barrier-based exclusivity.
The policy delineates a sharp boundary between text processing capabilities and resource-intensive modalities. File uploads, images, voice interactions, and image generation retain separate limits that apply across all tiers, meaning the volume expansion is isolated strictly to text-based messaging. This constraint confirms the infrastructure logic behind the shift: OpenAI is willing to subsidize unlimited text workloads where inference costs are predictable and manageable, while maintaining scarcity management on multimodal features that demand heavier compute overhead or expensive external model routing. The retention of these specific limits protects margins by preserving controls over the most variable cost drivers in the system.
The strategic context for this move rests heavily on OpenAI’s scale and market position. With weekly active users now exceeding 1 billion, the marginal cost of serving text-only queries at this magnitude is manageable enough to justify removing friction as a retention imperative. By eliminating text rate limits and introducing stronger default models like Luna across the free tier, OpenAI positions unlimited access as a defensive baseline against competitors who might compete on similar terms. The preservation of paid-tier controls for Sol and reasoning sliders reinforces that monetization is no longer about gating access but about offering reliability and customization to users who require them. This recalibration suggests OpenAI views platform dominance as dependent on the breadth of frictionless interaction rather than the restriction of core capabilities.