Home
ChatGPT Free Version Gets Major Upgrade With Epic GPT-5.6 Luna Boost Plus and Pro Users Gain Exclusive Perks
OpenAI has rolled out a significant update to ChatGPT today. Free-tier users now have GPT-5.6Luna set as their default model, and the Chat mode has removed text conversation limits, enabling unlimited usage. This change is being deployed to all free users throughout the week.
Plus and Pro subscribers have also received enhancements. GPT-5.6Sol, integrated into Chat, has undergone targeted optimization, with official statements highlighting marked improvements in factual accuracy and response quality. Altman announced on X: "5.6Sol performs significantly better in chats, and free users now enjoy unlimited text conversations!"
OpenAI reports that one billion people weekly use ChatGPT for quick Q&A, web searches, planning, research, consulting, and complex decision-making. This GPT-5.6Sol update is engineered to support this entire spectrum of use cases. The updated model delivers more focused responses, dynamically adjusting detail levels based on the query, avoiding unnecessary formatting, and providing constructive corrections when responses merely agree without adding value. For Plus and Pro users, the same model powers both instant answers and deep reasoning, ensuring a more consistent user experience.

Specifically, for quick questions, the model provides direct answers with essential context; for complex tasks like multi-step planning, research, or writing, it offers more comprehensive responses while keeping core recommendations clear. In OpenAI’s comparison examples, GPT-5.5Instant lists all data points—such as rain, wind, and temperature—in a cycling consultation scenario, whereas GPT-5.6Sol first addresses the user’s primary concern, identifies wind as the key factor rather than rain, retains only cyclist-relevant details, and even generates a direct weather forecast sheet.
Hallucination rates have also decreased significantly. Internal OpenAI assessments indicate that the proportion of prompts containing at least one factual error in financial, medical, and legal domains dropped by approximately 62% for GPT-5.6Luna and by about 68% for GPT-5.6Sol, compared to GPT-5.5Instant.
Subscribers also benefit from UI improvements. Following the release of the three 5.6 models, users were required to select both the model and inference strength simultaneously, which was cumbersome. The Chat interface now introduces a slider feature similar to that in Codex. Plus and Pro users can adjust the model’s reasoning depth across web, mobile, and desktop versions—keeping it in fast mode for daily queries and increasing the slider for planning, research, writing, coding, or decision-making.
Free users now have full access to GPT-5.6Luna with unlimited text chatting. However, features like image generation and file uploads remain restricted. Since Luna is not a universal model, OpenAI has added a new "Think" button to handle more complex issues.
Notably, this GPT-5.6Sol update applies only to the ChatGPT chat experience; the GPT-5.6Sol version used in Work and Codex is not included in this update.
Related article
Etched Successfully Secures Funding, Valuation Hits $5 Billion
Amidst the rapid evolution of artificial intelligence, AI inference chip startup Etched has just closed a funding round at a $5 billion valuation, while announcing annual sales hitting $1 billion. This milestone has energized the global AI chip secto
Anthropic Adjusts Claude Fable5 Access, Revoking Rights for Some Users Amid Expansion
Anthropic has updated its subscription access policy for the Claude Fable5 model. Beginning July 20, subscribers on Max and Team Premium plans retain access, though usage caps are significantly tightened. Post-reward period, standard quotas drop by 3
Harvey Launches Tenet Powered by Chinese Open-Source Kimi K3
American legal AI unicorn Harvey has unveiled Harvey Tenet, a specialized legal model built on Moonshot AI’s open-weight Kimi K3 through advanced post-training techniques. This milestone marks the first entry of a Chinese open-weight model into the t
Related Special Topic Recommendations
Comments (0)
0/500
OpenAI has rolled out a significant update to ChatGPT today. Free-tier users now have GPT-5.6Luna set as their default model, and the Chat mode has removed text conversation limits, enabling unlimited usage. This change is being deployed to all free users throughout the week.
Plus and Pro subscribers have also received enhancements. GPT-5.6Sol, integrated into Chat, has undergone targeted optimization, with official statements highlighting marked improvements in factual accuracy and response quality. Altman announced on X: "5.6Sol performs significantly better in chats, and free users now enjoy unlimited text conversations!"
OpenAI reports that one billion people weekly use ChatGPT for quick Q&A, web searches, planning, research, consulting, and complex decision-making. This GPT-5.6Sol update is engineered to support this entire spectrum of use cases. The updated model delivers more focused responses, dynamically adjusting detail levels based on the query, avoiding unnecessary formatting, and providing constructive corrections when responses merely agree without adding value. For Plus and Pro users, the same model powers both instant answers and deep reasoning, ensuring a more consistent user experience.

Specifically, for quick questions, the model provides direct answers with essential context; for complex tasks like multi-step planning, research, or writing, it offers more comprehensive responses while keeping core recommendations clear. In OpenAI’s comparison examples, GPT-5.5Instant lists all data points—such as rain, wind, and temperature—in a cycling consultation scenario, whereas GPT-5.6Sol first addresses the user’s primary concern, identifies wind as the key factor rather than rain, retains only cyclist-relevant details, and even generates a direct weather forecast sheet.
Hallucination rates have also decreased significantly. Internal OpenAI assessments indicate that the proportion of prompts containing at least one factual error in financial, medical, and legal domains dropped by approximately 62% for GPT-5.6Luna and by about 68% for GPT-5.6Sol, compared to GPT-5.5Instant.
Subscribers also benefit from UI improvements. Following the release of the three 5.6 models, users were required to select both the model and inference strength simultaneously, which was cumbersome. The Chat interface now introduces a slider feature similar to that in Codex. Plus and Pro users can adjust the model’s reasoning depth across web, mobile, and desktop versions—keeping it in fast mode for daily queries and increasing the slider for planning, research, writing, coding, or decision-making.
Free users now have full access to GPT-5.6Luna with unlimited text chatting. However, features like image generation and file uploads remain restricted. Since Luna is not a universal model, OpenAI has added a new "Think" button to handle more complex issues.
Notably, this GPT-5.6Sol update applies only to the ChatGPT chat experience; the GPT-5.6Sol version used in Work and Codex is not included in this update.
Etched Successfully Secures Funding, Valuation Hits $5 Billion
Amidst the rapid evolution of artificial intelligence, AI inference chip startup Etched has just closed a funding round at a $5 billion valuation, while announcing annual sales hitting $1 billion. This milestone has energized the global AI chip secto
Anthropic Adjusts Claude Fable5 Access, Revoking Rights for Some Users Amid Expansion
Anthropic has updated its subscription access policy for the Claude Fable5 model. Beginning July 20, subscribers on Max and Team Premium plans retain access, though usage caps are significantly tightened. Post-reward period, standard quotas drop by 3
Harvey Launches Tenet Powered by Chinese Open-Source Kimi K3
American legal AI unicorn Harvey has unveiled Harvey Tenet, a specialized legal model built on Moonshot AI’s open-weight Kimi K3 through advanced post-training techniques. This milestone marks the first entry of a Chinese open-weight model into the t











