OpenAI Removes ChatGPT Text Caps as It Deploys GPT-5.6 Models to 1 Billion Users
The AI developer is lifting messaging limits for standard chats while introducing new GPT-5.6 Luna and Sol models with enhanced reasoning controls.
OpenAI has eliminated messaging restrictions on text-based ChatGPT conversations for all free and entry-tier users, deploying a new default model line called GPT-5.6 as the platform reaches 1 billion weekly active users.
The artificial intelligence developer announced the overhaul alongside internal evaluation data showing a marked drop in hallucinations. Factual errors decreased 62% in the new base model, GPT-5.6 Luna, and 68% in the upgraded GPT-5.6 Sol model compared to the previous GPT-5.5-Instant engine baseline.
Under the infrastructure change, GPT-5.6 Luna will serve as the default model for Free and Go account holders, replacing GPT-5.5. Non-paying accounts are also getting a dedicated “Think” button to trigger deeper reasoning on technical or complex queries. OpenAI noted that while text interaction caps are gone, separate limits remain for file uploads, voice chats, images, and image generation.
Subscribers on ChatGPT Plus and Pro plans are receiving an upgraded GPT-5.6 Sol model tuned for rapid research, planning, writing, and concise decision-making. These higher tiers include a new “thinking slider” that allows users to adjust how much computational processing time the model spends on a problem based on its complexity. OpenAI emphasized that this model is distinct from the GPT-5.6 Sol build used in Codex and Work tools.
The upgraded GPT-5.6 Sol model rolled out to Plus and Pro accounts today, while rollout steps for Free and Go tiers will occur throughout the week. Access to unlimited text messaging and the “Think” button will be fully enabled by next week.
