Elon Musk’s xAI launches Grok 4 alongside a $300 monthly subscription
San Francisco, CA – Elon Musk’s artificial intelligence company, xAI, has officially unveiled its latest and most powerful AI model, Grok 4, marking a significant advancement in the competitive AI landscape. This launch includes an ultra-premium subscription plan, SuperGrok Heavy, priced at $300 per month, positioning xAI with the most expensive offering among major AI providers.
Grok, xAI’s answer to leading models like OpenAI’s ChatGPT and Google’s Gemini, is engineered to analyze images and provide comprehensive responses to user queries. In recent months, Grok has become deeply integrated into Musk’s social network, X (formerly Twitter), following xAI’s acquisition of the platform. This integration has, however, also brought Grok’s occasional missteps into the public eye for millions of users.
The introduction of Grok 4 comes with high expectations, particularly as it prepares to compete directly with OpenAI’s anticipated GPT-5 model, which is expected to launch later this summer.
During a livestream event on Wednesday night, Elon Musk asserted the new model’s prowess: “With respect to academic questions, Grok 4 is better than PhD level in every subject, no exceptions.” He further added, with characteristic candor, that while it may currently “lack common sense” and has not yet “invented new technologies or discovered new physics,” he believes “that is just a matter of time.”
This major announcement follows a turbulent period for Musk’s ventures, including the recent departure of Linda Yaccarino as CEO of X and prior incidents involving Grok’s automated account generating controversial, antisemitic comments. While these events were largely sidestepped during the Grok 4 launch, the focus remained firmly on the new models’ performance and capabilities.
xAI has released two distinct versions: Grok 4 and the more advanced, multi-agent Grok 4 Heavy. Musk explained that Grok 4 Heavy operates by spawning multiple AI agents to collaboratively tackle problems, comparing their findings “like a study group” to arrive at the optimal solution.
Initial benchmarks suggest Grok 4 is a formidable contender. On Humanity’s Last Exam—a rigorous test assessing AI’s proficiency across various subjects—Grok 4 (without tools) achieved an impressive 25.4%, surpassing Google’s Gemini 2.5 Pro (21.6%) and OpenAI’s o3 (21%). With tools, Grok 4 Heavy escalated its performance to 44.4%, significantly outperforming Gemini 2.5 Pro with tools (26.9%).
Furthermore, the nonprofit Arc Prize reported that Grok has set a new state-of-the-art score on its ARC-AGI-2 test—a challenging visual pattern recognition benchmark—with a score of 16.2%. This performance is nearly double that of the next best commercial AI model, Claude Opus 4.
The SuperGrok Heavy subscription, priced at a steep $300 per month, offers subscribers early access to Grok 4 Heavy and upcoming features. This premium tier aligns with similar offerings from industry leaders but notably stands as the most expensive in the current market. Subscribers can also look forward to an AI coding model in August, a multi-modal agent in September, and a video generation model in October.
In a strategic move to foster broader adoption, xAI is making Grok 4 available through its API, inviting developers to integrate the model into their applications. Despite its enterprise sector being nascent, only two months old, xAI intends to collaborate with hyperscalers to expand Grok’s availability across various cloud platforms.
While Grok 4 demonstrates frontier-level performance on benchmarks, xAI faces the challenge of overcoming recent controversies to establish Grok as a trusted and viable solution for businesses against established giants like ChatGPT, Claude, and Gemini. The market will closely watch whether enterprises are prepared to embrace Grok, acknowledging its cutting-edge capabilities alongside its past incidents.
