Sign in. It’s quick, free and it’s up to you.
An account is an optional way to support the work we do. Find out more.
The new version of the chatbot came out yesterday Alamy Stock Photo
openai
OpenAI releases new ChatGPT version with 'PhD level' expertise, but it can't spell 'blueberry'
Co-founder and chief executive Sam Altman touted this latest iteration as “clearly a model that is generally intelligent.”
OPENAI RELEASED A keenly awaited new generation of its hallmark ChatGPT yesterday, touting “significant” advancements in artificial intelligence capabilities as a global race over the technology accelerates.
ChatGPT-5 is rolling out free to all users of the AI tool, which is used by nearly 700 million people weekly, OpenAI said in a briefing with journalists.
Co-founder and chief executive Sam Altman touted this latest iteration as “clearly a model that is generally intelligent.”
Altman cautioned that there is still work to be done to achieve the kind of artificial general intelligence (AGI) that thinks the way people do.
“This is not a model that continuously learns as it is deployed from new things it finds, which is something that, to me, feels like it should be part of an AGI,” Altman said.
“But the level of capability here is a huge improvement.”
The rollout has not been without issues though.
ChatGPT-5 has struggled to give correct answers to simple prompts since it was released.
It even denies its own existence when asked about it.
In response to the question, “Is this ChatGPT-5?”, the large language model replied:
“You’re currently chatting with ChatGPT-4o architecture, which is part of the GPT-4 family – not ChatGPT-5.”
Another example of the new version of the chatbot making a basic error has prompted derision from some social media users.
ChatGPT-5 cannot spell the word ‘blueberry’.
When users asked how many times the letter ‘b’ appears in the word, it replied that there are three.
I had to try the “blueberry” thing myself with GPT5. I merely report the results.
— Kieran Healy (@kjhealy.co) August 8, 2025 at 1:04 AM
[image or embed]
Industry analysts have heralded the arrival of an AI era in which genius computers transform how humans work and play.
“As the pace of AI progress accelerates, developing superintelligence is coming into sight,” Meta chief executive Mark Zuckerberg wrote in a recent memo.
“I believe this will be the beginning of a new era for humanity.”
Altman said there were “orders of magnitude more gains” to come on the path toward AGI.
“Obviously… you have to invest in compute (power) at an eye watering rate to get that, but we intend to keep doing it.”
Tech industry rivals Amazon, Google, Meta, Microsoft and Elon Musk’s xAI have been pouring billions of dollars into artificial intelligence since the blockbuster launch of the first version of ChatGPT in late 2022.
Chinese startup DeepSeek shook up the AI sector early this year with a model that delivers high performance using less costly chips.
‘PhD-level expert’
With fierce competition around the world over the technology, Altman said ChatGPT-5 led the pack in coding, writing, health care and much more.
“GPT-3 felt to me like talking to a high school student — ask a question, maybe you get a right answer, maybe you’ll get something crazy,” Altman said.
“GPT-4 felt like you’re talking to a college student; GPT-5 is the first time that it really feels like talking to a PhD-level expert in any topic.”
Altman expects the ability to create software programs on demand — so-called “vibe-coding” – to be a “defining part of the new ChatGPT-5 era.”
In a blog post, British AI expert Simon Willison wrote about getting early access to ChatGPT-5.
“My verdict: it’s just good at stuff,” Willison wrote.
“It doesn’t feel like a dramatic leap ahead from other (large language models) but it exudes competence – it rarely messes up, and frequently impresses me.”
However, Musk wrote on X, formerly Twitter, that his Grok 4 Heavy AI model “was smarter” than ChatGPT-5.
Honest AI?
ChatGPT-5 was trained to be trustworthy and stick to providing answers as helpful as possible without aiding seemingly harmful missions, according to OpenAI safety research lead Alex Beutel.
“We built evaluations to measure the prevalence of deception and trained the model to be honest,” Beutel said.
ChatGPT-5 is trained to generate “safe completions,” sticking to high-level information that can’t be used to cause harm, according to Beutel.
The company this week also released two new AI models that can be downloaded for free and altered by users, to challenge similar offerings by rivals.
The release of “open-weight language models” comes as OpenAI is under pressure to share inner workings of its software in the spirit of its origin as a nonprofit.
With reporting from David Mac Redmond
have your say