xAI, Elon Musk’s AI company, has unveiled Grok 4 and Grok 4 Heavy, along with a new $300-per-month “SuperGrok Heavy” subscription, aiming to rival OpenAI’s ChatGPT and Google’s Gemini.
Grok 4 has demonstrated strong benchmark performance, scoring 25.4% on Humanity’s Last Exam without tools, surpassing Google’s Gemini 2.5 Pro (21.6%) and OpenAI’s o3 (21%). Musk stated that Grok 4 “is better than PhD level in every subject, no exceptions,” with respect to academic questions.
Grok 4 Heavy, which utilizes tools, achieved a score of 44.4%, outperforming Gemini 2.5 Pro (26.9%). Additionally, Grok set a new state-of-the-art score of 16.2% on the ARC-AGI-2 test, nearly doubling Claude Opus 4’s score.
The launch of Grok 4 follows a turbulent week for Musk’s companies, marked by Linda Yaccarino’s departure as X CEO and an incident where Grok’s automated X account made antisemitic comments. xAI briefly limited the account and removed the offensive posts.
Despite these events, xAI focused on Grok 4’s capabilities, announcing plans to release an AI coding model in August, a multi-modal agent in September, and a video generation model in October.
xAI intends to make Grok available through its API and cloud platforms, targeting developers and hyper-scalers.
