Musk Releases Grok 4! Tops All Benchmarks

Musk unveils Grok 4 AI model achieving postdoctoral-level performance across all subjects, topping benchmarks with 10x improved reasoning capabilities.

"AI Disruption" Publication 7100 Subscriptions 20% Discount Offer Link.

---

All subjects are at postdoctoral level.

After long anticipation, xAI's next-generation large model—Grok 4—has finally been released! Its capabilities exceed our imagination.

Today, the xAI launch event we've been eagerly awaiting finally began, with Musk appearing in the livestream. He opened by saying, "This is the world's best AI, let us show you."

Musk stated that Grok 4 achieves perfect scores on the SAT (American college entrance exam) every time without prior access to the questions, and can achieve near-perfect scores on the GRE in any subject, surpassing the level of all graduate students worldwide. Grok 4's most powerful aspect is its reasoning ability, which has achieved superhuman reasoning levels.

Musk believes that Grok 4 can achieve new scientific discoveries within this year.

Thanks to enhanced computational power and reinforcement learning training, Grok 4's reasoning ability has improved 10-fold compared to its predecessor. From Grok 2 to Grok 4, different technical paradigms were adopted: next token prediction, pretraining computation, pretraining + RL, and RL computation, respectively.