OpenAI has rolled out a bunch of new updates for ChatGPT over the past week after launching its latest GPT-5 model, bringing new models, expanded features, and upcoming personality changes to the AI assistant. The new updates touch both individual subscribers and enterprise users. So, in this article, I have compiled all the features announced by OpenAI in the past week.
1. GPT-4o Now Listed Under Legacy Models by Default
The GPT-4o is now available under “Legacy models” by default for paid users. This makes it easier for those who want to revisit the older but still capable model without extra setup. This is an important update as the company reintroduced it after removing it for a while. For users who were comfortable with GPT-4o’s style and performance, the change ensures it remains just a click away.
2. Additional Models Can Be Enabled in Settings
Now, the paid users get an option to toggle on “Show additional models” in their settings. It unlocks access to more specialized models, including o3, GPT-4.1, and GPT-5 Thinking mini. This gives power users more flexibility in choosing the right model for different tasks, whether that’s lightweight processing, deeper reasoning, or testing how newer models stack up against the older ones.
3. GPT-5 Now Has Auto, Fast, and Thinking Modes
Another functional update is that GPT-5 now comes with three modes in the model picker. The Fast mode is built for quick responses when speed is the priority, while Thinking takes more time to generate deeper and more structured answers. For those who don’t want to choose, the Auto option automatically switches between Fast and Thinking depending on the complexity of the query, giving users a balance between speed and depth.
4. More Messages on GPT-5 Thinking
The company has also expanded capacity for its heaviest users. Plus and Team subscribers can now use up to 3,000 messages per week on GPT-5 Thinking. Once that limit is reached, they will automatically get access to additional capacity on GPT-5 Thinking mini.
5. GPT-5 Expands to Enterprise and Education Users
Until recently, GPT-5 was only available to individual Plus and Pro subscribers. Now, Enterprise and Education customers have also got access to GPT-5. This means employees and students will be able to use the newest model with its faster, deeper, and more adaptive responses.
6. Gmail and Google Calendar Integration
Now Plus and Pro users are able to connect Gmail and Google Calendar to ChatGPT. With this integration, the assistant can pull in information from your inbox and schedule to provide more contextually useful responses. For example, ChatGPT can help draft replies to emails, remind you of upcoming events, or even generate summaries based on your calendar.
7. GPT-5 Is Getting Warmer and Friendlier
Along with the technical upgrades, OpenAI is also fine-tuning how GPT-5 communicates. The company says feedback showed that GPT-5 felt too formal to users, so it’s now being adjusted to sound warmer and more approachable.
Importantly, OpenAI stresses that these changes are not designed to make the model overly flattering or sycophantic. Internal tests show its behavior remains consistent, just less rigid. The rollout may take up to a day, and more personality adjustments are expected in the future.
Moreover, recently it was reported that the company has launched a new ChatGPT Go plan, priced at ₹399 per month, which is around $4. The company has also unveiled a new Voice speed feature, which also helps avoid repeating your custom instructions when using voice mode. However, there is no official confirmation about these updates from OpenAI as of now.
