After promising to launch Grok 3 as "the smartest AI model on earth," Elon Musk unveiled the highly anticipated model at a live event that attracted the attention of AI enthusiasts from around the world. The event came as major companies strive to deliver smarter and more efficient tools, and at a time when China's DeepSeek-R1 is well deservedly leading the scene.

Details of the unveiling of Grok 3 At 8 pm Pacific time, or 6 a.m. today, February 18, 2025, Egypt time, the Grok 3 model was launched amid an atmosphere of enthusiasm and anticipation on platform X. Elon Musk described his new model as the smartest on Earth, stressing that it offers significant progress compared to the previous version Grok 2, and this is not only in terms of computational ability, but in its ability to interact with an understanding that resembles human thinking, beyond expectations, and to show part of The model's capabilities were used during the live show to solve complex physical problems such as drawing planetary paths, highlighting its potential to deal even with unusual scientific challenges.

Grok 3 amazes Lex Friedman Not only did the excitement go live, but what made matters even more interesting came when MIT AI scientist and popular podcast host Lex Fridman wrote on his X account: I had the opportunity to use Grok 3 early and it really amazed me (he literally said "it's the dawn of my mind"). Such a statement, from someone like Lex Friedman, was only proof that the model exceeds the expectations of users and experts alike, but you should know that Elon and Lex are friends and have had more than one meeting before, so does Lex "compliment" his friend? May be!

Available Plans Grok 3 initially offers exclusive access to Premium Plus subscribers on platform X, with a new subscription model called "Super Grok" aimed at tech enthusiasts who want to experience advanced features first. It was also revealed that the voice interaction feature was added, which is a step to make handling Grok 3 more natural and smooth, allowing users to communicate with the system as if they were talking to a real person, just like with the "Advanced Voice" feature in ChatGPT.

What is Grok 3's mode relative to competitors? The Grok 3 model is designed to compete with the toughest models on the scene such as the o1-pro, and even the anticipated GPT-5, from OpenAI, as well as the most advanced models from Google (such as Gemini 2.0), Claude 3.5 Opus, Llama 3.1, DeepSeek, and others.

In the benchmark tests, Grok 3 demonstrated amazing abilities in logic, problem solving, and programming. The developers claim that their model has the upper hand over the rest of the models in performance metrics, exceeding or at least equaling the options mentioned in performing complex tasks and computational efficiency.

However, we should know that despite these internal tests conducted by xAI, which indicate the superiority of Grok 3, these claims still need to be independently evaluated. Some skeptical voices suggest that it is necessary to monitor how the model behaves.