Grok 4: Elon Musk's Latest AI Model Pushes Beyond All Previous Boundaries

Elon Musk's xAI unveiled its latest and most powerful artificial intelligence model, Grok 4, late Wednesday evening, along with a new SuperGrok Heavy subscription priced at $300 per month. Grok is xAI's answer to models such as OpenAI's ChatGPT and Google's Gemini, and it can analyze images and answer questions. In recent months, Grok has become more deeply integrated into Musk's social network X, which was recently acquired by xAI.

Breakthrough Benchmark Results

Grok 4 achieves state-of-the-art results that place it at the top among current AI models. According to data released by xAI, the model achieved the following scores:

Humanity's Last Exam is a particularly challenging test that measures an AI's ability to answer thousands of questions across a wide range of fields, including mathematics, the humanities, and the natural sciences. The nonprofit organization Arc Prize confirmed that Grok achieved a new record score on the ARC-AGI-2 test, which consists of puzzle-like problems requiring the identification of visual patterns.

"When it comes to academic questions, Grok 4 is better than PhD level in every subject, with no exceptions," Elon Musk said during Wednesday evening's livestream. "It may occasionally lack common sense, and it has not yet invented new technologies or discovered new physics, but that is only a matter of time."

Two Model Versions and a New Subscription

xAI launched two versions of the model: the standard Grok 4 and the more powerful Grok 4 Heavy, which is a "multi-agent version" with enhanced performance. According to Musk, Grok 4 Heavy creates several agents that work on a problem simultaneously and then compare their work "like a study group" to find the best answer.

The SuperGrok Heavy subscription, priced at $300 per month, offers:

• Early access to Grok 4 Heavy
• Early access to new features
• Priority support

This plan is similar to the ultra-premium tiers offered by OpenAI, Google, and Anthropic, but xAI now offers the most expensive subscription among major AI providers.

A Turbulent Week for Musk's Companies

The launch of Grok 4 comes amid a turbulent week for Elon Musk's companies. Earlier on Wednesday, Linda Yaccarino stepped down from her role as CEO of X after roughly two years at the company. X has not yet announced her successor.

Yaccarino's departure came just days after Grok's official automated account on X responded to users with antisemitic comments criticizing Hollywood's "Jewish executives" and praising Hitler. xAI had to temporarily restrict Grok's account and delete the offensive posts. In response to the incident, xAI apparently removed a recently added section from Grok's public system instructions that told it not to shy away from "politically incorrect" claims.

Musk and xAI's leadership largely avoided discussing the incident and instead focused on Grok 4's performance and capabilities.

Technical Advances and Speed of Development

The charts presented demonstrate the "absurd" pace of progress, as Musk called it. Grok 4 represents a tenfold improvement over Grok 3, which itself was a tenfold leap from Grok 2. The model benefits from a massive 100-fold increase in training capacity compared to the previous version, using the purpose-built Colossus supercomputer with 200,000 H100 GPUs.

Future Plans

xAI unveiled an ambitious roadmap for the coming months:

• August 2025: Coding model
• September 2025: Multimodal agent
• October 2025: Video generation model

SuperGrok Heavy subscribers may receive early access to some of the new products that xAI plans to launch in the coming months.

Integration and Availability

xAI is releasing Grok 4 through its API in an effort to encourage developers to build applications using the model. The company notes that xAI's enterprise division is only two months old, but it plans to work with hyperscalers (large cloud service providers) to make Grok available through their cloud platforms.

Expectations for Grok 4 are high. xAI's latest AI model will be compared with OpenAI's upcoming GPT-5 AI model, which is expected to launch later this summer.

Despite Grok's state-of-the-art benchmark performance, xAI may find it difficult to overcome its recent missteps as it attempts to present Grok to businesses as a genuine competitor to ChatGPT, Claude, and Gemini. Only time will tell whether businesses are ready to embrace Grok with all its shortcomings.