Following a major operational milestone of reaching one billion weekly active users, OpenAI is rolling out a comprehensive suite of updates across both its free and premium subscription tiers. The primary highlight of the update is the removal of text-based chat limits for standard accounts, significantly expanding standard access to advanced generative capabilities.

Driving the upgraded experience for free and lower-tier accounts is the new GPT-5.6 Luna model, which replaces the outgoing GPT-5.5 standard. In addition to unrestricted text messaging, non-paying users will now have access to a dedicated “Think” function, which allows users to explicitly request deeper computational reasoning for intricate queries. However, resource caps remain in effect for non-text interactions, including file processing, voice interaction, and image generation.

Subscribers on the Plus and Pro tiers are also receiving targeted upgrades centered on the new GPT-5.6 Sol engine, designed specifically for rapid responses, online research, planning, and drafting. Premium accounts will feature an adjustable reasoning slider, giving power users direct control over how much analytical effort the AI devotes to solving a given prompt.

Internal benchmarking indicates a substantial reduction in model hallucinations and inaccuracy. Official evaluations demonstrate that factual errors dropped by 62% on GPT-5.6 Luna and 68% on GPT-5.6 Sol when compared to earlier GPT-5.5 instant builds. The premium updates are being made available immediately, with free-tier enhancements expected to complete rollout over the coming week.

Key Takeaways

• Free ChatGPT users now have access to unlimited text-based conversations powered by the new GPT-5.6 Luna model.
• Plus and Pro subscribers gain the upgraded GPT-5.6 Sol engine along with a customizable reasoning slider to control computational effort.
• Internal testing demonstrates up to a 68% decrease in factual errors compared to previous model iterations.

Editor’s Analysis & Impact

OpenAI’s decision to grant unlimited text access to free users comes right as the platform crosses one billion weekly active users, representing a clear strategy to lock in market dominance. By lifting text restrictions and deploying the more accurate GPT-5.6 Luna model globally, OpenAI raises the baseline standard for free consumer AI tools, putting immense competitive pressure on rivals such as Google and Anthropic. Meanwhile, the introduction of features like the adjustable thinking slider highlights an emerging industry trend: allowing users to explicitly choose between low-latency speed and high-compute analytical depth. Furthermore, driving down factual errors by over 60% addresses the core issue of reliability, which remains the single biggest barrier to mainstream AI adoption in enterprise and research contexts.

Frequently Asked Questions

Q: Are all features now unlimited for free ChatGPT accounts?
A: No. While text-based conversations are now unlimited, usage limits remain in place for non-text features, including file analysis, voice mode, and image generation.

Q: What is the new 'Think' button in ChatGPT?
A: The 'Think' feature allows users to trigger enhanced analytical reasoning capabilities within the model when submitting complex or multi-step queries.

Q: How much more accurate are the new GPT-5.6 models?
A: Internal testing showed a 62% reduction in factual errors for GPT-5.6 Luna and a 68% reduction for GPT-5.6 Sol compared to the previous GPT-5.5-Instant model.