Elon Musk’s Grok 4 AI has been introduced with the bold claim of delivering top-tier performance in academic question-answering.
xAI, Elon Musk’s artificial intelligence startup, has officially launched Grok 4, along with its most expensive subscription plan to date, SuperGrok Heavy, priced at $300 per month.
Grok is positioned as a direct competitor to ChatGPT and Gemini, featuring capabilities like image analysis and question answering.
According to Musk, Grok 4 outperforms PhD-level expertise across all academic subjects without exception.
Alongside Grok 4, xAI also released Grok 4 Heavy, a more powerful and multimodal version of the model.
xAI claims that Grok 4 delivers performance on par with cutting-edge models in benchmarks such as Humanity’s Last Exam—a test consisting of thousands of questions in the humanities, mathematics, and natural sciences. Grok 4 scored 25.4% without using external tools, outperforming Gemini 2.5 Pro at 21.6% and OpenAI’s o3 (high) at 21%.
The Grok 4 Heavy model, with access to additional tools, achieved 44.4% in the same benchmark. By comparison, Gemini 2.5 Pro with tools reached 26.9%.
In another independent benchmark, ARC-AGI-2—which focuses on visual pattern recognition and puzzle solving—Grok scored 16.2%, nearly twice the performance of the runner-up, Claude Opus 4.
Grok 4 is now available via API, enabling developers to build applications on top of it. xAI is reportedly in talks with major cloud service providers to bring Grok to their cloud platforms.
The SuperGrok Heavy subscription not only offers early access to Grok 4 Heavy but also allows users to test upcoming features. This plan is currently the most expensive subscription among all major AI model providers.
The launch of Grok 4 comes during a turbulent week for Musk-led companies. Just hours before the announcement, Linda Yaccarino stepped down as CEO of X after nearly two years. A replacement has yet to be named.
Despite Grok’s technological advances, these recent developments may pose significant challenges to its adoption, especially when compared to more established competitors like ChatGPT, Claude, and Gemini. It remains unclear whether enterprise clients will be willing to trust Grok despite its current limitations.
