Elon Musk says his xAI company’s latest AI model Grok 4 is “better than PhD level in everything”.
Describing the latest model as “the smartest AI in the world” at its launch, the billionaire SpaceX and Tesla boss said it had received 100 times more training than the Grok 2 version, which was replaced by Grok 3 in February.
“It is remarkable to see how quickly artificial intelligence is evolving,” said Mr Musk, saying the new model would achieve near-perfect results in graduate exams in nearly every subject and expected the first “watchable half hour” of television produced by AI by the end of the year.
“Grok 4 is smarter than nearly all graduate students in all subjects simultaneously.”
Describing the current time as the “intelligence big bang”, he admitted Grok 4 “may lack common sense” but it might create new technology “as soon as this year”.
“The most important thing for any AI is to be truth-seeking… You can instil the right values to be truthful, honourable and the values you want to instil in a child you hope to grow up to be incredibly powerful,” he said.
The announcement comes in the wake of antisemitic commentary from the Grok 3 chatbot, which included praise of Adolf Hitler.
On Wednesday, the Grok’s X account posted: “We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts.”
Some experts have said Mr Musk has attempted to steer away from other chatbots such as OpenAI’s ChatGPT and Google’s Gemini, which is considered “woke”.
In June, he invited X users to help train the chatbot with “divisive facts” which he described as “things that are politically incorrect, but nonetheless factually true.”
On Wednesday, Linda Yaccarino said she was stepping down two years after Mr Musk hired her to run X, formerly Twitter, which he bought for 44 billion dollars (£32.4 billion) in late 2022.
