OpenAI ends chat limits for free ChatGPT users, rolls out faster GPT-5.6 models

OpenAI is removing conversation restrictions from ChatGPT's free tier, allowing unrestricted text-based chatting for all users. The shift comes as the platform recently surpassed 1 billion weekly active users.

New models and reasoning features#

The update introduces GPT-5.6 Luna, which will become the default model for free and Go tier subscribers, succeeding the earlier GPT-5.5. A fresh "Think" button is also rolling out to these users, enabling them to request deeper reasoning for complex queries.

Paid subscribers—both Plus and Pro tier members—gain access to a separate variant called GPT-5.6 Sol, optimized for quick tasks like research, decision-making, and content creation. This version comes with a thinking slider that lets users dial up or down the depth of analysis depending on the problem's complexity. OpenAI clarified that this upgrade applies only to standard chat use; the version powering Codex and Work tools remains unchanged.

Accuracy improvements and timing#

Internal testing shows notable gains in reliability across the board. GPT-5.6 Luna produces factual errors at a rate 62% lower than GPT-5.5-Instant, while GPT-5.6 Sol cuts errors by 68% in the same comparison.

The upgraded Sol model for Plus and Pro users is available immediately. Free and Go users will receive their new default model and Think button this week, with unlimited text conversations arriving by next week. File uploads, image handling, voice interactions, and image generation will continue to operate under existing rate limits.