OpenAI has released its newest model, GPT-5, bringing a major upgrade to its ChatGPT tool used by around 800 million people worldwide.
This new version promises big gains in areas like coding, writing, and task handling, as the company moves closer to developing artificial general intelligence (AGI) — a system that can perform most jobs better than humans.
According to OpenAI, GPT-5 is more capable at producing creative content and writing code for websites and apps.
It also reduces factual mistakes and no longer simply blocks unsafe questions.
Instead, it gives helpful answers within its safety rules or clearly explains why it cannot respond.
The model has also improved in how it responds to users without being overly agreeable.
Another major change is seen in ChatGPT’s agent feature, which can now do tasks like booking restaurants or online shopping.
With permission, it can access services like Gmail, Google Calendar, and contacts to make this process easier.
GPT-5 also works across text, voice, and images, making it easier to use in different ways.
The model now performs better when responding to health-related questions and can alert users to serious concerns, though OpenAI reminds users that the chatbot is not a substitute for medical or mental health professionals.
Despite these upgrades, OpenAI says GPT-5 still falls short of true AGI.
One key limitation is that the model cannot learn new information continuously as it is used.
This ability, according to OpenAI’s leadership, would be a major part of reaching AGI.
The company also rolled out two open models this week. While a free version of GPT-5 is available, it comes with limits.
Users who subscribe to the $200 monthly Pro plan get full access.
Meanwhile, reports suggest OpenAI may soon reach a valuation of $500 billion, making it one of the most valuable AI firms globally.