AI Showdown: There is a new global ranking. GrokLogo appears on smartphones with Elon Musk and Xai seen in the background. Credit: Lucia Fdez, Shutterstock
Elon Musk called the Grok 4 the smartest AI Alive, but the global rankings have just fallen, and a real winner may surprise you. Who controls the bots?
Musk claimed it was smarter than the alumni, but the scoreboard tells a different story.
He called it a genius. The scoreboard called it the average. Elon Musk’s Shiny New Ai Bot, Grok 4, attends school in front of the entire world of technology. Oppenheimer More than Iron Man.
The new Off-Glock 4′ Musk, which declares smarter than almost every graduate student in all fields, is now facing a brutal dose of reality. UC Berkeley Chatbot Arena – fundamentally AI Smarts Premier League – Latest rankings have been removed. What do you guess? Grok didn’t create a top two.
Mask’s “The World’s Smartest AI” has arrived Third.
The table toppings were Google’s Gemini 2.5, followed by Openai’s GPT-4O and GPT-4.5. The Grok 4 is tied third. It’s a very decent effort if the PR team doesn’t rely on “the world’s smartest AI” across social media.
Be honest – bronze isn’t bad, and It’s a wOrk inProgression. But when you tell everyone your robot can surpass Oxford, your robot finished third behind the regular suspect.
What is Grok? And why does Elon moan?
Grok is Musk’s answer to ChatGpt. It’s an edgy, opinionated chatbot cooked by his AI startup Xai. It lives within X (formerly Twitter), and was pitched as a free-thinking, free-talking and fearless alternative, according to the “awakening” competition.
But it got off to a rocky start. Not long ago, Glock spitted anti-Semitic and racist content when prompted. Even mask fans thought this was loosening the screws. others look It’s a blatant media trick feed AI that says meaningful so that you can publish negative reports about Musk and his company.
II didn’t do t“Stop the T, take care.
Is the leaderboard legal or is it just a vibe?
Some experts have questioned the scoreboard itself. According to a terrible report by Cohere researchers, there are several in the chatbot arena Dangerous The practice behind scene, Like Private pre-test before ranking, score removal, and even model swap go public.
Meta was caught doing exactly that – sent a secret version of the Llama 4 model to compete. It’s AI on par with what looks like a job interview with twins actually Qualification.
If the system is defective, does Grok’s bronze mean anything? That dePends Above Who is You ask. But even in this chaotic competition, the best models continue to rise to the top.
What do real champions do differently?
Google’s Gemini 2.5 is not leaning forward. They are trained to process text, images, code, and more, not only repeating the fluffs of the internet, but also to reason like a scientist. Openai’s GPT-4O is well known for its smooth, human-like dialogue, but the GPT-4.5 packs some of the most striking problem-solving skills seen in previous models.
In contrast, Grok focuses on attitude Scholar, And that shows.
Musk made a bold claim. But once Also, reality It’s coming out short. Or is it AppeaRs.
Want more AI drama, technology tantrums, and brainy bots Behave badly? Stay tuned atlasgazette Technology.
more Spanish Living News.
more English news from around Spain.
