OpenAI has claimed to have taken a “significant step” towards artificial general intelligence (AGI) with the launch of its latest upgrade to ChatGPT, but has admitted there are still “many things” missing in its quest to create a system able to do humans’ jobs.
The startup said its GPT-5 model, the underlying technology that will power its breakthrough AI chatbot, represents a big upgrade on its predecessors in areas such as coding and creative writing – and is also a lot less sycophantic.
It said the upgrade was being made available to all of ChatGPT’s 700 million weekly users immediately.
Sam Altman, OpenAI’s chief executive, called the model a “significant step forward” to achieving the theoretical state of AGI, which the startup defines as a highly autonomous system that outperforms humans at most economically valuable work – or, in other words, can do their jobs.
However, Altman admitted GPT-5 had not reached that goal yet. “[It is] missing something quite important, many things quite important,” said Altman, pointing to the model’s inability to “continuously learn”.
Altman said GPT-5 was “generally intelligent” and a “significant step on the path to AGI” but had not reached it yet by most people’s definition.
“I think the way that most of us define AGI, we’re still missing something quite important, many things quite important. But one big one is … this is not a model that continuously learns as it’s deployed from things it finds, which is something that, to me, feels like it should be part of AGI,” he said, adding that, nonetheless, it was a “huge improvement” on its predecessors.
During a GPT-5 launch event on Thursday, Altman said this latest version of ChatGPT was like having “access to a PhD-level expert in your pocket”. He described the previous version of the chatbot as a college student, and the version before that as a high school student.
AGI’s theoretical capabilities, and the determination of heavily backed tech companies to achieve it, has led to predictions from AI executives that white-collar jobs – from lawyers to accountants, doctors and bankers – will be wiped out by looming advances in the technology. Dario Amodei, the boss of AI developer Anthropic, has warned the technology could replace half of all entry-level office jobs in the next five years.
OpenAI said the main improvements in GPT-5 included: fewer factual errors, or hallucinations; better software coding, allowing it to create functional websites and apps; increased capability at creative writing; and, rather than “refusing” a prompt that breaches its guidelines outright, the model will instead try to give the most helpful response possible within safety guidelines, or at least explain why it cannot help.
The agent feature in ChatGPT – which carries out tasks such as finding restaurant availability and shopping online – will also be able to access users’ Gmail, Google calendar and contacts, if given permission.
As with its predecessors, GPT-5 can generate voice, image and text and can deal with queries in those formats too.
The company on Thursday demoed how GPT-5 can write hundreds of lines of code in seconds and create tools such as a French-language learning programme. Staff said the technology is also writing less robotically and had the chatbot create a “more nuanced” eulogy. Altman said ChatGPT was a good tool for healthcare advice too, bringing on stage a woman who had been diagnosed with cancer last year to talk about how the chatbot helped her decide whether or not to get radiation therapy.
The company said the upgraded ChatGPT would be better at answering health-related questions and would be more proactive at “flagging potential concerns” – such as serious physical or mental illness, too.
The startup stressed the chatbot was not a replacement for professional help, amid concerns AI tools could worsen the condition of people vulnerable to psychosis.
Nick Turley, the head of ChatGPT at OpenAI, said the model showed “significant improvements on sycophancy”, after the startup’s admission this year that the chatbot’s most sophisticated model had become too agreeable and could distress users or make them uncomfortable.
The latest model release comes as tech firms pour multibillion-dollar sums into their attempts to achieve AGI. On Tuesday, Google’s AI unit outlined its latest step towards AGI by showcasing an unreleased “world model”, while last week Mark Zuckerberg, the CEO of Facebook parent Meta, said the development of superintelligence – another theoretical state of AI even more powerful than AGI – was “now in sight”.
Investors’ faith in the potential for further breakthroughs, and AI’s ability to transform modern economies, has also spurred a valuation boom in companies such as San Francisco-based OpenAI. It was reported on Wednesday that OpenAI is in early talks about a sale of shares held by current and former employees that would value it at $500bn, overtaking Elon Musk’s SpaceX.
Although OpenAI also released two open models this week and has a free version of ChatGPT, the company makes its revenues from charging for subscriptions to the most powerful versions of the chatbot and from integrating its models into businesses’ IT systems. Use of the free version of GPT-5-backed ChatGPT will be capped, while users on the $200-a-month Pro package will get unlimited access.