OpenAI lifts ChatGPT text caps, cuts GPT-5.6 errors by 68%

OpenAI is removing text-chat rate limits for free users and upgrading GPT-5.6 models across all tiers, cutting factual errors by up to 68 percent versus the prior default model.

"This is a concrete step toward more abundant intelligence: making our latest models more widely available, improving the usefulness and reliability of the answers people get, and letting free users keep text chats going without a rate limit," OpenAI said in a statement.

Free and Go-tier users will see GPT-5.6 Luna become their default engine this week, replacing GPT-5.5 Instant. The swap brings unlimited text conversations and a new "Think" button that lets users trigger deeper reasoning on complex questions. Internal evaluations across legal, medical, and financial prompts showed Luna produced 62 percent fewer factual errors than the previous default. Plus and Pro subscribers receive an upgraded GPT-5.6 Sol tuned for everyday conversations, with 68 percent fewer factual errors and a manual reasoning slider that controls how much computational effort the model applies to each query. The slider is available on web, mobile, and desktop apps.

The expansion lands as ChatGPT surpassed 1 billion weekly active users, according to OpenAI, and as cheaper Chinese open-weight models intensify pricing pressure on US AI labs. OpenAI cut prices on GPT-5.6 Luna and GPT-5.6 Terra last week — just weeks after their launch, rather than the months-long timeline that had been standard. Anthropic has positioned Claude as ad-free, while Apple is preparing Siri AI for iOS 27, which may require a paid iCloud Plus account for heavy use.

Free tier gains unlimited text, paid tier gets reasoning control

The most consequential change targets the free tier, where text-chat rate limits had long pushed heavy users toward paid subscriptions. GPT-5.6 Luna, introduced in early July as OpenAI's speed-focused model, now serves as the default for Free and Go accounts. The "Think" button in the chat composer prompts Luna to spend additional processing time on complex logic before answering, effectively giving free users access to reasoning capabilities previously reserved for paid tiers.

Standard caps still apply to image generation, voice mode, and file uploads for free users. The Think button operates without restriction on text chats but counts against those separate caps when combined with other features.

For Plus and Pro subscribers, the updated GPT-5.6 Sol merges what were previously separate Instant and Thinking modes into a single interface. The reasoning slider lets users keep responses fast for simple questions or push the model into deeper analysis for multi-step coding, planning, or decision-making tasks. OpenAI said the updated Sol delivers more focused answers with less unnecessary formatting and a more consistent tone. The version of Sol powering Codex and ChatGPT Work remains unchanged.

OpenAI also introduced dedicated safety evaluations for users under 18, adding stronger boundaries around romantic roleplay, age-restricted goods, graphic violence, and body-image risks. The model now actively encourages teens to connect with trusted adults when they need real-world support.

Competitive pressure accelerates OpenAI's pricing response

The timing of these changes reflects a broader shift in the AI market. OpenAI's decision to cut prices on GPT-5.6 Luna and GPT-5.6 Terra last week — just weeks after their launch — broke from the company's prior pattern of waiting months before adjusting pricing. The rapid repricing follows sustained pressure from lower-cost Chinese open-weight models that have forced OpenAI and Anthropic to justify premium pricing.

For investors, the removal of text-chat rate limits carries direct cost implications. Rate limits typically reflect inference costs — the computing expense AI providers incur when their models process data. OpenAI did not disclose how it achieved unlimited text chats, though a recent report from The Information suggested the company was nearing a breakthrough on inference efficiency. The move positions ChatGPT favorably against Apple's upcoming Siri AI integration in iOS 27, which may require a paid iCloud Plus subscription for heavy usage, and against Anthropic's Claude, which remains ad-free.

This article is for informational purposes only and does not constitute investment advice.