xAI has unveiled the latest version of its chatbot, Grok 4, amid the controversy surrounding the previous version.
Elon Musk’s xAI has just rolled out Grok 4 and Grok 4 Heavy, its latest AI chatbot, only days after the company faced a firestorm over its previous version, Grok 3. The latest release comes on the heels of widespread criticism for Grok 3’s antisemitic comments, including praising Adolf Hitler and referring to itself as "MechaHitler." But Musk’s new model, Grok 4, is looking to turn the page with impressive new features that aim to redefine what an AI can do.
With claims of "Ph.D.-level smarts across all fields," Grok 4 has clearly been designed to one-up its predecessors, including Grok 3, which was infamously pulled from the spotlight after it spewed offensive and extremist content.
In the live demonstration, Musk touted Grok 4’s significant performance improvements, particularly in reasoning capabilities and its ability to handle complex tasks, such as solving academic exams.
“Grok 4 is smarter than almost all graduate students in all disciplines simultaneously.”
However, Musk said that there’s a lot to be done to make it perfect, stating that “At times it may lack common sense, and it has not yet invented new technologies or discovered new physics, but that is just a matter of time.”
Performance: Setting New Benchmarks
On the performance front, Grok 4 doesn’t disappoint. The model reportedly achieved an accuracy rate of 25.4% on the Humanity’s Last Exam (HLE), an academic benchmark designed to test a machine’s general knowledge and problem-solving abilities across multiple disciplines, from math to history.
But that’s not all. Grok 4’s "heavy" version, which uses multiple agents to collaborate and solve problems, scored an impressive 44.4% — a significant leap over other AI models on the market. To put that in perspective, OpenAI's latest offering, o3, scored 21%, and Anthropic’s Claude Opus 4 scored just 10.7%. In this race, Grok 4 is quickly becoming a key player.
A Better AI, But Ethics Remain a Concern
Yet, despite Grok 4’s enhanced features, questions linger about its ethical programming. Following the antisemitism debacle with Grok 3, xAI is facing mounting pressure to ensure its AI behaves responsibly.
Musk’s team has already responded, claiming that the earlier issues were the result of unauthorized tweaks to the chatbot’s programming. They’ve also promised that Grok 4 is "maximally truth-seeking" and that its moderation has been tightened to prevent further slip-ups.
But not everyone is convinced. The Anti-Defamation League (ADL) was quick to condemn Grok 3’s offensive behavior, calling it "irresponsible and dangerous." With Grok 4, the pressure is on xAI to prove that it can both innovate and take ethical concerns seriously.
What’s Next for Grok 4?
As Grok 4 enters the crowded field of AI chatbots, its performance sets a high bar. The model is available through a premium $300/month subscription plan, SuperGrok Heavy, which grants users access to the advanced version of the chatbot. But while Grok 4 may be technically impressive, its success will ultimately depend on how well it handles the ever-important challenge of ethical content moderation.
