Elon Musk’s xAI has garnered significant attention with the release of Grok 3, its latest AI chatbot, receiving high marks from early users and experts.
Launched on February 17, 2025, during a livestreamed event in Los Angeles, Grok 3 has been described as “the smartest AI on Earth” by Musk and praised for its advanced reasoning and coding capabilities.
Early adopters have highlighted its conversational tone and cutting-edge performance, positioning it as a serious competitor in the crowded AI landscape. Users and industry figures have expressed enthusiasm for Grok 3’s capabilities.
Andrej Karpathy, former Tesla AI director and OpenAI co-founder, called it a “state-of-the-art thinking model.” He noted that it outperformed recent releases from DeepSeek and Google in complex tasks.
Ethan Mollick, an AI reviewer and Wharton professor, described it as “frontier-level” with impressive scaling but emphasized the need for independent testing to confirm its full potential.
Business leaders like Aaron Levie of Box praised xAI’s rapid development, crediting Musk’s ability to assemble top-tier talent and infrastructure. Grok 3 represents a massive leap in development compared to its predecessor. Just 18 months ago, xAI researchers considered the chatbot “almost like a toy.”
Grok 3 Pushes AI Boundaries
Now, it boasts significant computational advancements. Training Grok 3 required 200 million GPU-hours across 200,000 Nvidia H100 GPUs—ten times the compute of Grok 2. This infrastructure, built in just eight months and dubbed “Colossus,” enabled Grok 3 to excel in benchmarks.
It scored 52 on the AIME’24 math test, beating GPT-4o’s 48. It also achieved a record-breaking 1,400 ELO rating in Chatbot Arena and outperformed competitors on science and coding tests. Despite the praise, challenges remain. The AI industry lacks standardized benchmarks, making comparisons difficult.
Early reviews of “Deep Search,” a research tool launched alongside Grok 3, suggest it may lag behind competitors like OpenAI in depth and comprehensiveness. Experts also caution that without API access for independent evaluation, Grok 3’s claims remain provisional.
As xAI enters an increasingly competitive AI race against rivals like OpenAI and DeepSeek, early user enthusiasm underscores Grok 3’s potential to disrupt the field. However, its long-term success depends on sustained innovation and transparent validation of its capabilities.
This article was drafted with automated assistance and reviewed before publication. How we use AI · Report an error · Editorial responsibility: Matthias Camenzind, Editor-in-Chief