OpenAI improves GPT-5.6 Sol in ChatGPT and restricts free users to its weakest model

OpenAI Restricts Free ChatGPT Users to Its Weakest Model While Boosting GPT-5 for Subscribers

OpenAI has introduced significant changes to its ChatGPT models, including a major upgrade for paying users and a strict downgrade for free tier users. The company has enhanced GPT-5 and GPT-4.5 with improved “style” controls and reasoning capabilities, while simultaneously limiting free users to GPT-4o mini, its least powerful current model.

The shift aims to push more users toward paid subscriptions. Free users now only have access to the standard GPT-4o mini, losing the ability to use GPT-4o, the previous mainstay model.

The New Model Tiers Explained

Free Users Demoted to GPT-4o Mini

Free tier users can now only select the standard GPT-4o mini. This is the smallest and fastest model, but also the weakest in terms of reasoning and depth. The previous default GPT-4o is no longer available for free accounts.

Paid Subscribers Get GPT-5 and GPT-4.5 with “Style” Control

Plus and Pro subscribers will see a new model picker featuring GPT-5 and GPT-4.5. These premium models come with an updated “style” selector that allows users to adjust a model’s conversational tone, including options for more concise or more detailed responses.

How the New “Style” Selector Works

The style selector operates on a sliding scale between “conciseness” and “detail.” This replaces the old, separate model dropdowns for different reasoning levels (like “o3-mini” or “GPT-4o”).

Key Insight: The new system merges model selection with tone adjustment. A user can now pick “GPT-5” and then fine-tune whether it responds with short headlines or full essays.

OpenAI claims this reduces confusion for users who previously had to choose between distinct reasoning models for different tasks.

What This Means for Your Workflow

For Free Users: Lower Quality, Same Speed

If you are on the free tier, expect less accurate reasoning, shorter memory for context, and weaker performance on complex tasks. The trade-off is that GPT-4o mini remains fast and cheap for OpenAI to run, but the ceiling on output quality has been lowered.

For Paid Users: More Control, Higher Cost

Subscribers gain the ability to switch between GPT-5 and GPT-4.5, plus the granular style control. However, the GPT-5 model is reported to consume more tokens per query, meaning higher usage rates for heavy users on tiered plans.

The Strategic Shift Behind the Changes

OpenAI is clearly segmenting its user base. By reserving stronger reasoning models for paid tiers, the company creates a clearer value proposition for subscriptions. The timing aligns with increased competition from open-source models and rival closed-source products like Google’s Gemini and Anthropic’s Claude.

Critical Warning: Free users should expect a noticeable drop in response quality for multi-step reasoning, code generation, and analytical tasks. If you rely on ChatGPT for work, the free tier may no longer be viable.

What Happens Next

The model restrictions are rolling out over the next week. Free users will see their model picker automatically default to GPT-4o mini. Paid users will need to manually select the new GPT-5 option to gain access to the upgraded reasoning capabilities.

Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.

What are your thoughts on this? I’d love to hear about your own experiences in the comments below.