ChatGPT-5 is now live for all users, bringing a sharp jump in speed and skill, OpenAI said on Thursday.
The new model powers the world’s most popular chatbot, already used by nearly 700 million people each week.
ChatGPT-5 moves closer to human-level tasks
OpenAI chief Sam Altman called the upgrade “a clear step toward general intelligence.” In tests the bot handled coding, writing and medical queries better than any earlier version. The team says the model can act as an “agent,” running computer tasks on its own and even building simple apps on demand.
Developers showed the bot creating a French-learning app in seconds. They said ChatGPT-5 now answers like “a PhD-level expert” rather than a college student. The company trained the system to refuse unsafe requests and avoid deception. Safety lead Alex Beutel said new checks aim to keep answers useful but harmless.
The launch comes in a fierce global race. Google, Microsoft and Chinese labs are pouring billions into similar tools. OpenAI is also releasing lighter “open-weight” models for free, hoping to stay ahead while easing pressure to share its code.
Earlier this week the firm offered US federal workers a year of ChatGPT Enterprise for one dollar, seeking wider trials in government use. Altman said huge spending on computer power will continue as the company hunts for the next leap toward true artificial general intelligence.
For everyday users the change is simple. Open your account and the new ChatGPT-5 model is ready to chat, code and explain—no upgrade fee required.
