OpenAI refreshes ChatGPT with GPT-5.6 Sol and wider free access
OpenAI has updated GPT-5.6 Sol in ChatGPT for Plus and Pro subscribers and is expanding access for Free users. The company says the revised Sol model produces more focused responses and improved factual reliability, while a new control lets paid users select how much thought ChatGPT should put into an answer.
Free and Go users will receive GPT-5.6 Luna as the default model this week. Starting next week, OpenAI says they will gain unlimited text chats and a new Think button for questions that require deeper reasoning, subject to abuse guardrails. Limits for file uploads, images and other tools will remain in place.
Focused responses and adjustable effort
OpenAI describes GPT-5.6 Sol as a model tuned for the varied ways people use ChatGPT, from short questions and web searches to planning, research, writing and complex decisions. It is intended to adapt the level of detail to the request, use tighter formatting and avoid unnecessary material.
For Plus and Pro accounts, GPT-5.6 Sol now supports both Instant responses and deeper reasoning in the Chat experience. OpenAI says this is intended to make a move to higher effort feel like a more comprehensive response from the same assistant rather than a switch to a model with a different tone or behaviour.
A slider in ChatGPT for web, mobile and desktop lets those users select response effort. It can remain low for everyday questions or be increased for planning, research, writing, coding and decisions that need more thought. The release follows work on ChatGPT medical quality improvements, where factual quality in ChatGPT’s medical use was a central concern.
Internal reliability results
OpenAI reported results from an internal evaluation involving financial, medical and legal prompts that required factual detail. In that test, answers containing at least one factual error were about 62% less common with GPT-5.6 Luna and 68% less common with GPT-5.6 Sol than with GPT-5.5 Instant.
The company also says Sol is better at using sources it finds when responding to questions that depend on dates, numbers, rules, sources or assumptions. These statements describe OpenAI’s internal evaluation rather than an independently published benchmark.
Availability and safety boundaries
The updated GPT-5.6 Sol and its response-effort slider are available to Plus and Pro users from today. This particular version of Sol is limited to ChatGPT’s Chat experience; the GPT-5.6 Sol used by Work and Codex is unchanged in this release.
OpenAI says it added safety training and evaluations for users it believes are under 18. The training includes avoiding romantic roleplay, age-restricted challenges and presenting the model as a substitute for real-world relationships, along with boundaries around several sensitive content areas.
For businesses, the rollout makes it important to match a task to the available plan and capability: unlimited text conversations may support routine work, but tool limits and the unchanged Work and Codex models remain relevant when designing operational workflows.