OpenAI has just launched ChatGPT 5, and it could be the biggest update to the AI chatbot since its inception in 2022.
For over 700 million of us, ChatGPT has become a go-to tool for everything from answering questions and brainstorming ideas to helping with work, studies and daily life.
With the launch of GPT-5, OpenAI is promising a smarter, more reliable and more personal experience that aims to feel less like chatting with a robot and more like having a conversation with a real expert.
It’s not just about raw intelligence either; the company has also focused on making ChatGPT-5 more accurate, more customisable and easier to use, with new features designed to fit more comfortably into your daily routines. Here’s everything you can expect from the new and improved ChatGPT.
Smarter conversations with fewer errors
According to OpenAI’s CEO, Sam Altman, using ChatGPT 5 should feel more like talking to a real expert. He claims that this is the first time the chatbot “really feels like talking to a PhD-level expert on any topic.”
In reality, that means that you should expect more accurate answers, better explanations and far fewer AI hallucinations (those weird, made-up facts AI sometimes spits out) whenever you use ChatGPT.
OpenAI claims that GPT-5 is 26% less likely to make factual mistakes than GPT-4o, and that the new ‘thinking’ mode is even more reliable, with a 65% reduction in hallucinations versus the o3 model.
Personalisation and new features
ChatGPT 5 isn’t just about making the AI chatbot smarter – it should also be more personal than ever. You can now pick your chat colour to help differentiate between different styles of chat – personal, work, creation, etc – and even choose from four preset personalities: Cynic, Robot, Listener and Nerd.
These personalities are limited to the text-based version of ChatGPT 5 for now, but they’ll soon also be available in voice mode, so you can have conversations that match your mood or style.
It’s a massive improvement on what came before too; while you could tell ChatGPT to act or respond to questions, you had to delve deep into the chatbot’s settings and tell it specifically how you’d like it to respond. This adaptation is much simpler.
It’s not the only new ease-of-use feature either; no longer will you need to manually select the model you think is best required for the task at hand. Simply ask your query, and ChatGPT 5 will automatically pick the best model for your question.
Those who subscribe to the paid ChatGPT plans will also soon be able to connect Gmail, Google Contacts and Google Calendar directly into ChatGPT. The AI will automatically reference your emails, contacts and schedules whenever relevant, turning the chatbot into what could be a digital personal assistant.
Handles more and remembers more
ChatGPT 5 is said to feature a huge 256,000 token context window that OpenAI claims can keep track of much longer conversations, documents or code – ideal for work and creative purposes.
It essentially means that you can have more in-depth chats, upload bigger files or work on more complex projects without the AI getting confused and losing the thread.
What’s more, the model is designed to “fail gracefully” if it can’t answer a question, rather than make something up. OpenAI says it has been tested extensively to reduce the risk of harmful or misleading responses.
For everyday users, ChatGPT 5 is a massive leap forward that makes the AI chatbot more accurate, more personal and more helpful with its new features and renewed focus on reliability.
The new model is rolling out to both free and paid users of ChatGPT right now.
