ChatGPT’s New Model Picker: GPT-5 Options and the Return of GPT-4o
ChatGPT’s New Model Picker: GPT-5 Options and the Return of GPT-4o
ChatGPT users, rejoice! OpenAI is giving you more control over your AI experience. The latest update brings back the beloved GPT-4o and introduces a new model picker with several GPT-5 options. Let’s dive into what this means for you.
More Choices, More Control
In a recent announcement, OpenAI CEO Sam Altman revealed that ChatGPT now allows users to choose between different GPT-5 variants: Auto, Fast, and Thinking. This is a significant shift from the previous unified experience, offering users the ability to tailor the AI’s performance to their specific needs.
- Auto: This is the default option, designed to provide a balanced and efficient experience. It’s likely the best choice for most users.
- Fast: Need quick responses? The Fast version prioritizes speed, making it ideal for brainstorming or rapid information retrieval.
- Thinking: For more comprehensive and detailed answers, the Thinking version digs deeper and provides longer, more thoughtful responses. This is perfect for complex tasks and in-depth analysis.
The Return of GPT-4o
Many users were disappointed when OpenAI initially removed GPT-4o with the launch of GPT-5. The decision was met with criticism from those who preferred GPT-4o’s unique personality and capabilities. Good news: GPT-4o is back in the model picker for all paying users by default!
OpenAI listened to the feedback and has restored GPT-4o, ensuring that users can continue to leverage its strengths. Altman has also promised to provide ample notice before deprecating any models in the future.
Accessing Additional Models
Want even more options? You can toggle to “Show additional models” in ChatGPT’s web settings to access o3, 4.1, and GPT-5 Thinking mini. However, GPT-4.5 remains exclusive to subscribers of the $200-a-month Pro tier, as it requires significant GPU resources.
Rate Limits for GPT-5 Thinking
Keep in mind that GPT-5 Thinking now has a rate limit of 3,000 messages per week. Once you reach this limit, you’ll be switched to GPT-5 Thinking mini. This is a necessary measure to manage resource allocation and ensure a smooth experience for all users.
OpenAI Working on a Warmer Personality
OpenAI is also working on refining GPT-5’s personality to make it more approachable and engaging, while avoiding the perceived annoyances of GPT-4o. This ongoing effort aims to strike a balance between helpfulness and user-friendliness. According to Altman, they are striving for a personality that is warmer and more relatable.
Actionable Takeaway: Experiment with the New Model Picker
The best way to understand the differences between the GPT-5 variants is to try them out! Take some time to experiment with Auto, Fast, and Thinking to see which one best suits your specific needs and workflows. You might be surprised at how much of a difference the right model can make.
Expert Commentary (Simulated)
“The reintroduction of GPT-4o and the addition of multiple GPT-5 options demonstrate OpenAI’s commitment to user feedback and customization,” says Dr. Anya Sharma, an AI researcher at the fictional TechForward Institute. “This level of control empowers users to tailor the AI’s performance to their specific tasks, leading to more efficient and effective interactions.”
FAQ
Q: Why did OpenAI remove GPT-4o in the first place? A: OpenAI initially removed GPT-4o to simplify the ChatGPT experience with the launch of GPT-5, aiming for a unified model. However, user feedback led to its reintroduction.
Q: How do I access the different GPT-5 models? A: You can select your preferred GPT-5 model (Auto, Fast, or Thinking) through the model picker in the ChatGPT interface.
Q: What is GPT-5 Thinking mini? A: GPT-5 Thinking mini is a version of GPT-5 Thinking that you’ll be switched to once you exceed the 3,000 messages per week rate limit for the full GPT-5 Thinking model.
Q: Is GPT-4.5 available to all users? A: No, GPT-4.5 is currently only available to subscribers of the $200-a-month Pro tier.
Q: Will OpenAI deprecate GPT-4o again? A: Sam Altman has promised that OpenAI will provide ample notice before deprecating any models in the future.
Key Takeaways
- ChatGPT now offers multiple GPT-5 variants (Auto, Fast, Thinking) to customize your AI experience.
- GPT-4o is back for all paying users, addressing previous user concerns.
- Additional models (o3, 4.1, GPT-5 Thinking mini) are accessible through web settings.
- GPT-5 Thinking has a rate limit of 3,000 messages per week.
- Experiment with the different models to find the best fit for your needs.
Source: Engadget
