Breaking News
OpenAI is updating ChatGPT with changes aimed at improving everyday conversations while broadening access for users on its free plan, as the company continues refining how its models balance speed, depth, and accuracy.
For Plus and Pro subscribers, the update reworks GPT-5.6 Sol, the model powering ChatGPT's chat experience, to deliver more reliable factual answers and tighter, more focused responses. Alongside the update, OpenAI is rolling out a new slider that lets users control how much reasoning effort the model applies to each response. Free-tier users, meanwhile, are getting GPT-5.6 Luna as their new default model, along with unlimited text conversations and a "Think" button that unlocks deeper reasoning for more complex questions.
OpenAI says ChatGPT now serves roughly 1 billion people weekly across tasks ranging from simple lookups to research, planning, and complex decision-making. The updated GPT-5.6 Sol is designed to better handle that range - adjusting its level of detail based on the question, cutting unnecessary formatting, and pushing back with a correction rather than simply agreeing when that wouldn't actually help the user. For Plus and Pro subscribers, the same underlying model now handles both instant replies and more involved reasoning tasks, a shift OpenAI says creates a more unified experience. According to the company, the model is built to keep answers focused - offering quick, direct replies for simple questions while preserving fuller detail for more complex tasks like multi-step planning or research, without burying the core recommendation.
OpenAI says the update also targets accuracy, particularly in cases involving dates, figures, sourcing, or rules-based reasoning, by making better use of retrieved sources. In internal testing on financial, medical, and legal prompts requiring precise factual accuracy, the company found that responses containing at least one factual error dropped by about 62% with GPT-5.6 Luna and 68% with GPT-5.6 Sol, compared with the earlier GPT-5.5 Instant model.
Part of the update focuses on making ChatGPT's fast and slow response modes feel more consistent with each other. OpenAI says the goal is for switching from an instant reply to a higher-effort response to feel like the same model spending more time thinking, rather than switching to an entirely different system with its own tone. The new slider — available to Plus and Pro users on web, mobile, and desktop — lets people dial reasoning effort up or down depending on the task, from routine questions to more demanding work like coding or research.
Free users are getting broader access as part of the rollout, including unlimited text chats powered by GPT-5.6 Luna. A new Think button will also let free users request deeper reasoning on harder questions, giving the model more time to work through complex prompts.
The updated GPT-5.6 Sol and the new effort slider are available immediately to Plus and Pro users. GPT-5.6 Luna will become the default model for Free and Go users this week, with unlimited text chats and the Think button following next week, subject to abuse safeguards. Limits will remain in place for file uploads, image generation, and other tools. OpenAI noted that this version of GPT-5.6 Sol is built specifically for everyday conversational use and will only be available in ChatGPT's Chat experience — the version powering OpenAI's Work and Codex products is not changing under this release.
For Plus and Pro subscribers, the update reworks GPT-5.6 Sol, the model powering ChatGPT's chat experience, to deliver more reliable factual answers and tighter, more focused responses. Alongside the update, OpenAI is rolling out a new slider that lets users control how much reasoning effort the model applies to each response. Free-tier users, meanwhile, are getting GPT-5.6 Luna as their new default model, along with unlimited text conversations and a "Think" button that unlocks deeper reasoning for more complex questions.
OpenAI says ChatGPT now serves roughly 1 billion people weekly across tasks ranging from simple lookups to research, planning, and complex decision-making. The updated GPT-5.6 Sol is designed to better handle that range - adjusting its level of detail based on the question, cutting unnecessary formatting, and pushing back with a correction rather than simply agreeing when that wouldn't actually help the user. For Plus and Pro subscribers, the same underlying model now handles both instant replies and more involved reasoning tasks, a shift OpenAI says creates a more unified experience. According to the company, the model is built to keep answers focused - offering quick, direct replies for simple questions while preserving fuller detail for more complex tasks like multi-step planning or research, without burying the core recommendation.
OpenAI says the update also targets accuracy, particularly in cases involving dates, figures, sourcing, or rules-based reasoning, by making better use of retrieved sources. In internal testing on financial, medical, and legal prompts requiring precise factual accuracy, the company found that responses containing at least one factual error dropped by about 62% with GPT-5.6 Luna and 68% with GPT-5.6 Sol, compared with the earlier GPT-5.5 Instant model.
Part of the update focuses on making ChatGPT's fast and slow response modes feel more consistent with each other. OpenAI says the goal is for switching from an instant reply to a higher-effort response to feel like the same model spending more time thinking, rather than switching to an entirely different system with its own tone. The new slider — available to Plus and Pro users on web, mobile, and desktop — lets people dial reasoning effort up or down depending on the task, from routine questions to more demanding work like coding or research.
Free users are getting broader access as part of the rollout, including unlimited text chats powered by GPT-5.6 Luna. A new Think button will also let free users request deeper reasoning on harder questions, giving the model more time to work through complex prompts.
The updated GPT-5.6 Sol and the new effort slider are available immediately to Plus and Pro users. GPT-5.6 Luna will become the default model for Free and Go users this week, with unlimited text chats and the Think button following next week, subject to abuse safeguards. Limits will remain in place for file uploads, image generation, and other tools. OpenAI noted that this version of GPT-5.6 Sol is built specifically for everyday conversational use and will only be available in ChatGPT's Chat experience — the version powering OpenAI's Work and Codex products is not changing under this release.
See What’s Next in Tech With the Fast Forward Newsletter
SECURITY
Tweets From @varindiamag
Nothing to see here - yet
When they Tweet, their Tweets will show up here.
