ChatGPT gets a major upgrade across all tiers as far as its reasoning abilities are concerned. OpenAI announced that it is dropping daily text-chat limits for non-paying users, while simultaneously rolling out upgraded AI models and granular reasoning controls across every subscription tier, utilising the latest GPT-5.6 models.
The update comes shortly after ChatGPT crossed a milestone of 1 billion weekly active users. OpenAI CEO Sam Altman took to his social media feed on X (formerly Twitter) to announce the update, stating, “5.6 Sol much better in chat now, and unlimited text chat for free users!”
For context, the GPT-5.6 model family was released a few weeks ago as OpenAI’s most advanced models, moving towards a multi-tier capability with high token efficiency in focus. The 5.6 model is divided into three distinct variants:
– Sol is the flagship engine built for complex reasoning, advanced coding, and cybersecurity.
– Terra is a balanced, cost-effective workhorse
– Luna is the lightweight, ultra-fast model.
ChatGPT upgrades for users: What changes across tiers
With the update, OpenAI brings a couple of changes across various ChatGPT subscription tiers to enhance user retention. The tiers in question include Free, Go, Plus, and Pro, with no limits for text queries anymore. Users also gets more control over reasoning and unique new Think button.
Let’s take a look at what each of the ChatGPT tiers will now offer.
Free and ChatGPT Go users
Unlimited text chats: The daily message caps on standard text conversations have been completely removed. Users can engage in uninterrupted long text conversations without hitting rate-limit barriers.
New default model is GPT-5.6 Luna: With the update, OpenAI is replacing the previous default GPT-5.5 Instant with the newer GPT-5.6 Luna – this becomes the primary engine for Free and Go tiers. OpenAI reports its internal evaluations showed Luna reducing factual errors by 62 per cent compared to its predecessor. This should allow for better usability while researching.
The ‘Think’ button: Free and Go users also gain access to a dedicated ‘Think’ toggle, allowing them to manually allocate additional processing time on a per-prompt basis for multi-step reasoning, complex math, or structured problem-solving. Consider this an easier access to more processing for research topics.
Limits still remain: Note that heavy computing tasks, including image generation (DALL·E), file uploads, data analysis, and voice modes, still retain their standard fair-use quotas. Only the text conversations are now unlimited.
ChatGPT Plus and Pro subscribers
Upgraded core model: Paid subscribers of ChatGPT Plus and Pro receive an enhanced version of GPT-5.6 Sol, tuned for high-speed, direct, and factually rigorous everyday tasks such as research, planning, and coding.
Interactive reasoning slider: Plus and Pro users also have a UI slider across web, mobile, and desktop interfaces. This control lets subscribers adjust the model’s reasoning effort, from immediate, concise answers to deep, extended analysis, thus offering more customisation with processing.
Better formatting and tone: The retuned Sol model avoids unnecessary fluff, provides tighter output formatting, and is claimed to proactively offer constructive corrections when user assumptions are incorrect. OpenAI says it won’t do passive agreement.