OpenAI rolls out GPT-5 for free and paid users; gets major upgrades in coding and reasoning

OpenAI has rolled out GPT-5, its most advanced artificial intelligence model yet, in a move to stay ahead in the fast-evolving AI landscape. The model, which was announced during a livestreamed event on Thursday, promises significant upgrades in reasoning, coding, and creative writing.

Wider access starting today

Starting today, GPT-5 will be available to all free users and many paying ChatGPT subscribers. Enterprise and education customers are expected to get access from next week. OpenAI says paid users will benefit from higher usage limits.

Described by CEO Sam Altman as a “major upgrade”, GPT-5 is designed to feel more like talking to an expert. The model responds faster, answers more accurately, and makes up information less often compared to earlier versions. It can also decide on its own when to take more time to process a question, instead of relying on the user to choose a specific AI mode.

Features of GPT-5

As part of the launch, OpenAI has introduced four new chatbot personalities, Cynic, Robot, Listener and Nerd, which are currently available as a “research preview”. These are meant to let users customise how ChatGPT interacts with them.

A key focus of GPT-5 is on software development. OpenAI demonstrated how the model can quickly build a fully functional web app with games and interactive features, based on simple instructions. Early testers, including startups like Lovable and Anysphere, said the model is noticeably better at generating, fixing, and maintaining code.

OpenAI is placing a strong emphasis on GPT-5’s usefulness for software development. With more tech companies and startups building AI tools to assist developers in writing and debugging code more efficiently, a new wave of “vibe coding” is taking shape. 

Sam Altman described this shift as central to GPT-5’s impact, saying that the concept of “software on demand” will define this next phase of AI.

The release comes at a time when OpenAI is facing growing pressure from competitors in both the US and China. Companies like Google, Anthropic, xAI (founded by Elon Musk), and DeepSeek are all pushing forward with their own advanced models. Meta, meanwhile, has been expanding its AI team, reportedly hiring several OpenAI staff.

To maintain its edge, OpenAI has secured tens of billions in funding to support the development of its models, data centres and computing infrastructure. It is also in early discussions for a stock sale that could value the company at around $500 billion, according to Bloomberg.