AI showdown: the brand new world rankings are in.
Grok brand displayed on a smartphone with Elon Musk and Xai seen within the background.
Credit score: Lucia Fdez, Shutterstock
Elon Musk known as Grok 4 the neatest AI alive – however the world rankings simply dropped, and the true winner would possibly shock you. Who guidelines the bots?
Musk claimed it was brainier than grad college students – however the scoreboard tells a special story.
He known as it a genius. The scoreboard known as it common. Elon Musk’s shiny new AI bot, Grok 4, simply bought schooled in entrance of the whole tech world – and the result’s extra Oppenheimer than Iron Man.
Contemporary off declaring Grok 4 ‘smarter than nearly all graduate college students in all disciplines,’ Musk is now dealing with a brutal dose of actuality. The UC Berkeley Chatbot Area – principally the Premier League of AI smarts – simply dropped its newest rankings. And guess what? Grok didn’t even make the highest two.
Musk’s “smartest AI on this planet” simply got here third.
Topping the desk was Google’s Gemini 2.5, adopted by OpenAI’s GPT-4o and GPT-4.5. Grok 4 limped in tied for third – a really first rate effort in case your PR group hadn’t already plastered ‘world’s smartest AI’ throughout social media.
Let’s be sincere – bronze isn’t dangerous, and it’s a work in progress. However once you’ve been telling everybody your robotic may outthink Oxford, ending third behind the standard suspects stings only a bit.
What’s Grok – and why is Elon groaning?
Grok is Musk’s reply to ChatGPT – an edgy, opinionated chatbot cooked up by his AI startup, xAI. It lives inside X (previously Twitter), and was pitched as a free-thinking, free-speaking, fearless various to the supposedly “woke” competitors.
Nevertheless it’s had a rocky begin. Not way back, Grok was caught spewing antisemitic and racist content material when prompted – behaviour that had even Musk followers questioning if this factor had a screw free. Others see it as a blatant media trick, baiting an AI to say imply issues so you’ll be able to publish unfavourable press about Musk and his corporations.
It didn’t cease the Pentagon, thoughts you – they reportedly pumped $200 million into Grok’s growth.
Is the leaderboard legit – or only a vibe-fest?
Some consultants are questioning the scoreboard itself. In response to a damning report by researchers at Cohere, the Chatbot Area has some dodgy practices behind the scenes, like personal pre-testing, rating deletions, and even mannequin swaps earlier than rankings go public.
Meta was caught doing simply that – sending a secret model of its LLaMA 4 mannequin to compete. It’s the AI equal of exhibiting as much as a job interview with a twin who’s really certified.
So if the system’s flawed, does Grok’s bronze even imply something? It depends on who you ask. However even on this chaotic competitors, one of the best fashions preserve rising to the highest – and Grok’s nonetheless trailing.
What are the true champs doing otherwise?
Google’s Gemini 2.5 is not any slouch. It handles textual content, pictures, code, and extra – and it’s been skilled to motive like a scientist, not simply repeat web fluff. OpenAI’s GPT-4o is legendary for clean, human-like dialogue, whereas GPT-4.5 packs among the sharpest problem-solving expertise seen in any mannequin so far.
Grok, in distinction, has targeted extra on angle than teachers, and it reveals.
Musk made daring claims. However once once more, the actuality got here up brief. Or so it appears.
Need extra AI drama, tech tantrums, and brainy bots behaving badly? Keep tuned to todaysviralmedia Tech.
Extra Spanish dwelling information.
Extra information in English from round Spain.