Elon Musk’s artificial intelligence startup, xAI, has unveiled Grok-3, the newest iteration of its chatbot. The launch comes at a critical juncture as the company seeks to establish itself in the competitive AI landscape, squaring off against firms like DeepSeek, OpenAI, and Google.
Grok-3 is designed to integrate with X, formerly known as Twitter. The chatbot is available immediately to Premium+ subscribers of X, and xAI has also introduced a new subscription tier, SuperGrok, for users accessing the chatbot via its mobile app and the Grok.com website. Grok-3 can generate text and images, and it doesn’t have the same restrictions on content as some competing models.
During a live stream alongside three xAI engineers, Musk stated that Grok-3 is “in a league of its own.” He claimed it surpasses its predecessor, Grok-2, with “more than 10 times” the computing power and achieving high scores in industry benchmark tests. The CEO characterized the bot as a tool for “maximally truth-seeking AI, even if that truth is sometimes at odds with what is politically correct.”
Musk pointed out that Grok-3 uses a “Big Brain” mode for research tasks. The latest version incorporates a smart search engine called DeepSearch, which xAI describes as a reasoning-based chatbot capable of explaining its thought process when responding to user inquiries. The tool offers functionalities for research, brainstorming, and data analysis.
Gil Luria, managing director at DA Davidson, remarked that “The introduction of Grok-3 puts xAI back in the race for leadership in open-source LLMs [large language models]. It outperforms the current state-of-the-art models on some benchmarks, which makes xAI relevant again.”
As the race for AI supremacy intensifies, xAI is expanding its data center capabilities to train more advanced models, fueled by billions of dollars in fundraising. Musk highlights its supercomputer cluster in Memphis, Tennessee, known as “Colossus,” as the world’s largest.
Luria, however, noted that the improvements over the Grok-2 model may not fully justify the resources invested in its training.
