Summary
OpenAI CEO Sam Altman has announced major ChatGPT updates, introducing three GPT-5 response modes—Auto, Fast, and Thinking—along with a 196k-token reference limit for longer conversations. GPT-5 Thinking has a weekly 3,000-message cap, with GPT-5 Thinking Mini available afterward. GPT-4o has returned for paying users, and a “Show additional models” toggle now unlocks options like o3, 4.1, and GPT-5 Thinking Mini. GPT-4.5 remains Pro-only due to higher costs. Altman also revealed plans to make GPT-5’s personality more engaging, with GPT-5 touted as a major upgrade in coding, reasoning, and creative writing.
Table of Contents
Introduction
August 13, 2025 — OpenAI CEO Sam Altman has announced a significant round of updates to ChatGPT, giving paying users greater control over the company’s latest GPT-5 models and expanding the range of available AI options.
In a detailed post on X (formerly Twitter), Altman revealed new response modes, extended processing capabilities, and the return of popular models like GPT-4o. These changes are aimed at improving customization, flexibility, and user satisfaction for both casual and power users.
Three New Response Modes for GPT-5
The biggest highlight is the introduction of three distinct response modes for GPT-5:
Auto Mode – The default setting, optimized for balanced speed and reasoning.
Fast Mode – Prioritizes quicker responses for users who need instant answers.
Thinking Mode – Focuses on deeper, more detailed reasoning and complex problem-solving.
While Altman noted that Auto Mode will likely remain the most popular choice, the additional settings cater to users with specific needs—whether for rapid back-and-forth interactions or thorough, in-depth analysis.
Also Read: OpenAI has launched ChatGPT-5, the most advanced version of its AI chatbot
Limits and Expanded Capacity
For GPT-5 Thinking, paying users are now subject to a weekly rate limit of 3,000 messages. Once that limit is reached, users can continue with GPT-5 Thinking Mini, a scaled-down but still capable alternative.
In addition, GPT-5 now boasts an impressive 196k-token reference limit, allowing it to handle longer conversations, large documents, and extensive codebases in a single session. Altman hinted that these limits could evolve over time based on user demand and server capacity.
GPT-4o Returns to the Model Selector
One of the biggest surprises for paying customers is the return of GPT-4o to the model chooser. This model—known for its engaging and human-like tone—had been temporarily unavailable.
“If we ever turn it off, we’ll provide full details,” Altman assured users.
A new “Show additional models” toggle in ChatGPT’s web settings will also let paying users access experimental and specialized models, including o3, 4.1, and GPT-5 Thinking Mini.
However, the GPT-4.5 model will remain exclusive to Pro subscribers due to its higher GPU cost.
Personality Tweaks for GPT-5
Altman also revealed that OpenAI is working to make GPT-5’s personality “more engaging” without being overly intrusive—striking a balance between the professional tone of GPT-5 and the conversational style of GPT-4o.
“We don’t want it to be as annoying (to most users) as GPT-4o,” Altman joked.
GPT-5: More Capable Than Ever
Unveiled during a livestream event last Thursday, GPT-5 is described by OpenAI as a massive leap in AI capabilities, excelling in:
Complex reasoning
Advanced coding
Creative writing
Expert-level problem-solving
In fact, OpenAI claims the chatbot can now pass for a PhD-level expert in various fields.
During a press briefing earlier this week, Altman called GPT-5 a “huge upgrade” over past models:
“For the first time, it feels like talking to an expert on a subject,” he said.
Why This Matters for Users
These updates show OpenAI’s commitment to customization and accessibility for different user needs. With multiple modes, larger memory limits, and more model options, paying customers now have unprecedented control over their ChatGPT experience.
For developers, researchers, and content creators, the extended token limit means fewer interruptions when working on large projects, while Fast Mode could be a game-changer for real-time collaboration.
Quick Summary of Key Updates:
Three response modes: Auto, Fast, Thinking.
Weekly limit: 3,000 messages for GPT-5 Thinking, with GPT-5 Thinking Mini as backup.
Extended context: 196k-token reference limit.
GPT-4o returns for all paying users.
Show additional models toggle for more AI options.
GPT-4.5 remains Pro-only.
Personality updates coming soon for GPT-5.
Final Thoughts
With these latest improvements, OpenAI is positioning GPT-5 as the most versatile AI assistant yet, giving users speed, depth, and flexibility all in one platform. Whether you’re looking for quick answers, detailed expert-level reasoning, or engaging conversations, the newest ChatGPT update ensures there’s a mode for you.
As AI continues to evolve, one thing is certain—OpenAI is determined to keep ChatGPT at the forefront of the AI revolution.
Check Here: Smart Solar Innovations: Exploring the Best Solar-Powered Products for a Greener Tomorrow