OpenAI is giving ChatGPT one of its biggest everyday overhauls yet. The company is removing text message caps for free accounts while rolling out an updated version of its flagship GPT-5.6 Sol model for ChatGPT paid subscribers. These changes aim to make responses faster, far more factual, and customizable.
Unlimited text chats and a new “Think” button for free accounts
The most impactful change for everyday users centers on free accessibility. OpenAI is replacing the older GPT-5.5 Instant model with GPT-5.6 Luna as the default engine for Free and entry-level Go tier accounts. Along with this model upgrade comes unlimited text-based conversations, removing the frustrating rate limits that previously halted long brainstorming sessions. Standard caps will still apply to image generation, voice mode, and file uploads.
Free users are also getting a new “Think” button built right into the chat composer. Tapping this option prompts GPT-5.6 Luna to spend extra processing time working through complex logic before answering. In internal testing across legal, medical, and financial prompts, GPT-5.6 Luna reduced factual errors by 62% compared to GPT-5.5 Instant.
Custom reasoning sliders and tighter formatting for paid tiers
ChatGPT Plus and Pro subscribers are getting a specially tuned version of GPT-5.6 Sol made for standard conversations. Instead of dumping walls of unnecessary formatting or filler context, the updated model focuses on concise, direct answers. On accuracy benchmarks, GPT-5.6 Sol produced 68% fewer factual errors than GPT-5.5 Instant. OpenAI clarified that this update strictly applies to general ChatGPT conversations, leaving the versions powering Codex and ChatGPT Work untouched.
Paid accounts are also getting a unified interface that merges Instant and Thinking modes into a single experience. A new manual slider—available on web, mobile, and desktop apps—lets users choose exactly how much reasoning effort the AI applies. Keeping the slider low delivers lightning-fast answers for simple questions, while pushing it up instructs the model to analyze complex coding, planning, or decision-making tasks deeply.
Enhanced safety guardrails and rollout timeline
Alongside model performance, OpenAI introduced dedicated safety evaluations for users under 18. The model now includes stronger boundaries around romantic roleplay, age-restricted goods, graphic violence, and body-image risks. Plus, it actively encourages teens to connect with trusted adults when they need real-world support.
The updated GPT-5.6 Sol model and manual reasoning slider are going live for Plus and Pro subscribers today. Free and Go users will see GPT-5.6 Luna become their default engine this week, with unlimited text chats and the “Think” button rolling out fully over the coming days.
