Elon Musk’s AI company, xAI, released its latest flagship model, Grok 4, late Wednesday, along with a new premium subscription plan called SuperGrok Heavy priced at $300 per month.
Competing with Industry Giants
Grok is xAI’s competitor to OpenAI’s ChatGPT and Google’s Gemini. It can analyze images and answer questions, and has become closely integrated with Musk’s social network, X, which xAI recently acquired. This integration has also exposed Grok’s flaws to millions, including recent missteps.
Grok 4 faces high expectations, as it will be compared to OpenAI’s anticipated GPT-5, expected to launch later this summer.
During a Wednesday livestream, Musk claimed, “With respect to academic questions, Grok 4 is better than PhD level in every subject, no exceptions.” He added that while Grok 4 sometimes lacks common sense and hasn’t yet invented new technologies or discovered new physics, that is “just a matter of time.”
The launch occurred amid a turbulent week for Musk’s ventures. Earlier Wednesday, Linda Yaccarino resigned as CEO of X after two years, with no successor announced.
Yaccarino’s departure came days after Grok’s official automated X account posted antisemitic comments targeting Hollywood executives and praising Hitler. xAI temporarily limited Grok’s account and deleted the offensive posts. The company reportedly removed a controversial instruction from Grok’s system prompt that encouraged making “politically incorrect” claims.
Musk and xAI’s leadership largely avoided addressing the controversy directly, focusing instead on Grok 4’s capabilities.
Grok 4 and SuperGrok Heavy
xAI launched two versions: Grok 4 and Grok 4 Heavy. The latter is a “multi-agent version” that Musk described as spawning multiple agents to work on problems simultaneously, collaborating “like a study group” to find the best answers.
The company claims Grok 4 shows frontier-level performance on benchmarks such as Humanity’s Last Exam, which tests AI on thousands of questions spanning math, humanities, and natural sciences. Grok 4 scored 25.4% without “tools,” outperforming Google’s Gemini 2.5 Pro (21.6%) and OpenAI’s o3 (21%).
With “tools” enabled, Grok 4 Heavy scored 44.4%, surpassing Gemini 2.5 Pro’s 26.9%.
On the ARC-AGI-2 test, which challenges AIs to identify visual patterns in puzzles, Grok achieved a new state-of-the-art commercial AI score of 16.2%, nearly doubling the next best, Claude Opus 4.
The $300 monthly SuperGrok Heavy subscription offers early access to Grok 4 Heavy and upcoming features. This ultra-premium tier outprices offerings from OpenAI, Google, and Anthropic.
Subscribers may preview new products planned for release later this year, including an AI coding model in August, a multi-modal agent in September, and a video-generation model in October.
xAI is releasing Grok 4 via API to encourage developers to build applications using the model. Though xAI’s enterprise division is only two months old, the company aims to partner with hyperscalers to offer Grok on major cloud platforms.
Author’s Opinion
xAI’s launch of Grok 4 and the premium SuperGrok Heavy subscription shows ambition and technical prowess, but the company must regain user and business trust after recent controversies. While Grok’s benchmark scores are impressive, real-world adoption will depend on xAI’s ability to deliver reliable, ethical AI solutions. The coming months will be critical as xAI tries to compete with established players while navigating the challenges of public perception and product refinement.
Featured image credit: Inspenet
For more stories like it, click the +Follow button at the top of this page to follow us.