OpenAI unveiled on Thursday its GPT-5 artificial intelligence model, the latest iteration of the technology behind ChatGPT, promising broader access and enhanced enterprise capabilities.
GPT-5 will be made available to all 700 million ChatGPT users, OpenAI said, as it seeks to demonstrate significant leaps in performance that justify continued heavy investment in AI infrastructure.
The release comes amid a surge in capital spending by the world’s leading technology firms Alphabet, Meta, Amazon and Microsoft, OpenAI’s principal backer which together plan to invest nearly US$400 billion in AI data centres this fiscal year.
“We’ve built GPT-5 to feel like you can ask a legitimate expert, a PhD-level expert, anything,” OpenAI Chief Executive Sam Altman told reporters at a San Francisco briefing. He highlighted the model’s prowess in software development “software on demand” as well as its strong performance in writing, health-related queries and financial analysis.
In demonstrations, GPT-5 generated complete, working programmes from plain-text prompts, a capability dubbed “vibe coding”. Early reviewers praised its coding skills and proficiency in science and mathematics but noted that the upgrade from GPT-4 to GPT-5 seemed more incremental than previous generational jumps.
Despite these improvements, Altman conceded that GPT-5 cannot yet learn autonomously, a milestone needed to match human adaptability. He noted that OpenAI’s “test-time compute” technology where the model applies extra processing power to tackle difficult questions will be accessible to the public for the first time, helping to address tasks that demand advanced reasoning.
Industry economists warn that enterprise spending on AI remains tepid compared with strong consumer uptake of applications like ChatGPT. “Consumer spending on AI just isn’t going to be nearly enough to justify all the money being spent on data centres,” said Noah Smith, an economics writer.
OpenAI is in preliminary talks to allow employees to liquidate holdings at a US$500 billion valuation, up from its current US$300 billion. Meanwhile, leading AI researchers command signing bonuses of up to US$100 million.
Nearly three years after ChatGPT’s debut ignited the generative AI boom, OpenAI’s challenge remains sustaining transformative leaps in capability as it confronts a data-scarcity barrier and the technical complexity of training ever-larger models.
“Building more infrastructure globally is essential,” Altman said, underscoring the need for widespread, localised AI availability to fulfil the company’s mission of benefiting all humanity.
