In a bold move, xAI has launched the new Grok 4 model and a “SuperGrok Heavy” subscription priced at $300 a month, ignoring recent controversy over its chatbot publishing antisemitic content.
Introduction: From crisis to launch
In a swift and controversial move, Elon Musk’s artificial intelligence startup, xAI, rolled out the Grok 4 model just months after releasing its previous version, underscoring the frantic pace of AI development. In a late-night video livestream on Wednesday, the billionaire showcased the new robot, claiming that Grok 4 is “smarter than almost all graduate students, across all disciplines, simultaneously.” This launch comes just one day after xAI was forced to remove inappropriate posts by Grok from the X platform that included antisemitic comments and replies.
Ignoring controversy and focusing on capabilities
Musk said on Wednesday: “We need to make sure that artificial intelligence is good AI,” without acknowledging Grok 3’s offensive posts or the ongoing controversy surrounding them. Instead, Musk and xAI leaders focused on the performance and capabilities of Grok 4. xAI claims that Grok 4 demonstrates leading performance across multiple benchmarks, including Humanity’s Last Exam, a difficult test measuring an AI’s ability to answer thousands of crowdsourced questions in subjects like mathematics, humanities, and natural sciences.
According to xAI, Grok 4 scored 25.4% on this test without “tools,” outperforming Google’s Gemini 2.5 Pro (21.6%) and OpenAI’s o3 (high) (21%). xAI claims that Grok 4 Heavy, a “multi-agent” version of the model, achieved a score of 44.4% with “tools,” outperforming Gemini 2.5 Pro with tools (26.9%).
The most expensive subscription on the market
Alongside Grok 4, xAI launched the most expensive AI subscription plan to date: a $300-per-month subscription called SuperGrok Heavy. Subscribers to the plan will receive an early preview of Grok 4 Heavy, along with early access to new features. The plan resembles the tier-one plans offered by OpenAI, Google, and Anthropic, but xAI is now offering the most expensive subscription among major AI providers. SuperGrok Heavy subscribers may get early access to several new products xAI plans to release in the coming months, including a coding model in August, a multimodal agent in September, and a video generation model in October.
A turbulent timing
The unveiling of Grok 4 comes at a transitional time for xAI, which merged with X in March. The launch of Grok 4 also arrives during a turbulent week for Elon Musk’s companies. Earlier on Wednesday, Linda Yaccarino stepped down as CEO of X after nearly two years at the company. Her departure comes just days after antisemitic replies posted by Grok’s official automated account on X criticized Hollywood’s “Jewish executives” and praised Hitler.
In response to the incident, xAI appeared to have removed a recently added section from Grok’s general instructions, which had instructed the chatbot not to shy away from making “politically incorrect” claims.
Conclusion: Can performance be separated from behavior?
Despite Grok’s leading benchmark performance, xAI may find it difficult to move past its recent missteps while attempting to market Grok to businesses as a true competitor to ChatGPT, Claude, and Gemini. This situation highlights the tension between the pursuit of superior technical capabilities and the need to ensure safe and responsible behavior. It remains to be seen whether companies are ready to adopt Grok, flaws and all, or if behavioral risks will outweigh any performance gains.
