OpenAI has announced a major expansion of ChatGPT access, rolling out its updated GPT-5.6 models to Free and Go tier users while removing text chat limits altogether.
The update, part of the company’s broader push to make advanced AI “benefit all of humanity,” introduces two distinct models tailored to different user needs: GPT-5.6 Sol for Plus and Pro subscribers, and GPT-5.6 Luna for Free and Go users.
For Plus and Pro users, GPT-5.6 Sol now unifies Instant and deep reasoning responses into a single consistent model, eliminating the jarring shift in tone that previously occurred when switching between quick answers and thinking mode.
A new slider lets Plus and Pro users manually control how much computational effort ChatGPT dedicates to each query, ranging from fast everyday answers to more thorough analysis for coding, planning, and research tasks.
OpenAI says the model now delivers tighter, more focused responses, cuts unnecessary formatting, and pushes back with corrections rather than blindly agreeing when a user’s assumption is incorrect.
Accuracy improvements headline the release. In internal testing across financial, medical, and legal prompts, factual errors dropped by roughly 62 percent with GPT-5.6 Luna and 68 percent with GPT-5.6 Sol compared to the previous GPT-5.5 Instant model, addressing a persistent criticism of large language models around hallucinated facts, dates, and figures.
Free and Go users stand to gain the most from this rollout. GPT-5.6 Luna becomes the default model for these tiers this week, and starting next week, unlimited text chats will be available without the rate limits that previously capped usage.
A new Think button will also let free users trigger deeper reasoning for harder questions, though this remains subject to anti-abuse guardrails. Limits on file uploads, image generation, and other tools will still apply.
OpenAI also detailed expanded safety measures in its accompanying system card, particularly protections for users believed to be under 18, including restrictions on romantic roleplay, age-restricted content, and prompts encouraging real-world support connections when needed.
Notably, GPT-5.6 Sol’s chat-optimized version won’t extend to OpenAI’s Work or Codex products, which retain their current model configuration. This update signals OpenAI’s continued strategy of democratizing access to frontier AI capabilities while tightening factual reliability across its consumer product line.
Strengthen Your SOC by Accelerating Threat Detection & Rapid Investigations. -> Integrate ANY.RUN With Your SOC Now.
