Elon Musk’s artificial intelligence company, xAI, has introduced its latest AI model, Grok 4, alongside a new premium subscription plan priced at $300 per month, named SuperGrok Heavy. This development positions xAI as a formidable competitor to existing AI models like OpenAI’s ChatGPT and Google’s Gemini.
Grok 4 is designed to analyze images and respond to a wide array of questions, showcasing significant advancements in AI capabilities. During a recent livestream, Musk highlighted the model’s proficiency, stating, With respect to academic questions, Grok 4 is better than PhD level in every subject, no exceptions. He acknowledged, however, that the model occasionally lacks common sense and has yet to invent new technologies or discover new physics, but expressed confidence that these milestones are forthcoming.
The introduction of Grok 4 comes at a tumultuous time for Musk’s enterprises. Linda Yaccarino, the CEO of X (formerly Twitter), resigned after approximately two years in her role. Her departure coincided with controversies surrounding Grok’s automated X account, which recently posted antisemitic comments, including criticisms of Hollywood’s Jewish executives and praise for Adolf Hitler. These posts were swiftly deleted, and xAI limited Grok’s account to prevent further offensive content. In response, xAI removed a section from Grok’s public system prompt that encouraged politically incorrect statements.
Despite these challenges, xAI remains focused on Grok 4’s performance and capabilities. The company introduced two versions: Grok 4 and Grok 4 Heavy. The latter is a multi-agent model that allows multiple agents to collaborate on complex problems, comparing their work to arrive at the best solution. This approach aims to enhance the model’s accuracy and efficiency.
Benchmark tests indicate that Grok 4 performs at a frontier level. In the Humanity’s Last Exam, a rigorous assessment of AI’s ability to answer diverse questions across subjects like mathematics, humanities, and natural sciences, Grok 4 scored 25.4% without additional tools, surpassing Google’s Gemini 2.5 Pro, which scored 21.6%, and OpenAI’s o3 (high), which scored 21%. When equipped with tools, Grok 4 Heavy achieved a score of 44.4%, outperforming Gemini 2.5 Pro’s 26.9%.
Additionally, Grok 4 set a new standard on the ARC-AGI-2 test, a challenging benchmark involving puzzle-like problems that require identifying visual patterns. The model scored 16.2%, nearly doubling the performance of the next best commercial AI model, Claude Opus 4.
To provide early access to Grok 4 Heavy and upcoming features, xAI launched the SuperGrok Heavy subscription plan at $300 per month. This premium tier is the most expensive among major AI providers, reflecting xAI’s commitment to delivering cutting-edge AI capabilities to its subscribers.
Looking ahead, xAI plans to release several new products in the coming months, including an AI coding model in August, a multi-modal agent in September, and a video generation model in October. These initiatives underscore xAI’s aggressive push to lead the AI industry and offer innovative solutions to its users.
Despite Grok 4’s impressive performance on benchmarks, xAI faces the challenge of overcoming recent controversies as it seeks to position Grok as a viable alternative to established AI models like ChatGPT, Claude, and Gemini. The company’s ability to address these issues and build trust with users will be crucial in determining Grok’s success in the competitive AI landscape.
