Elon Musk’s Grok 4 is smarter than PhD holders, but it might lack common sense!

Elon Musk’s artificial intelligence firm xAI launched its latest model, Grok 4, during a livestream on X (formerly Twitter). Though the event began an hour late, Musk introduced new features including multimodal input, faster reasoning, and improved voice interaction. “With respect to academic questions, Grok 4 is better than PhD level in every subject, no exceptions,” Musk claimed. Trained on xAI’s Colossus supercomputer, Grok 4 is positioned to outperform existing models such as GPT-3 and Gemini 2.5 Pro across key benchmarks. Musk suggested Grok 4 could go further still. “Reality is the ultimate reasoning test. We've run out of test questions to ask,” he said, predicting the AI could begin discovering new technologies by the end of 2025 and even 'new physics' within two years.

Musk further stated that Grok 4 surpasses PhD-level expertise across all academic disciplines, “without exception.” Despite this, Musk acknowledged the model 'may lack common sense' and has not yet achieved scientific breakthroughs.

From code interpreter to controversial replies

A standout feature of Grok 4 is Grok 4 Code, designed to write, debug, and explain software more efficiently. Musk encouraged users: “You can cut & paste your entire source code file into the query entry box on grok.com and Grok 4 will fix it for you!” This functionality mirrors tools such as GitHub Copilot and GPT-4 Code Interpreter, aiming to simplify tasks for developers by handling diagnostics and solutions automatically.

The model’s development, however, has not been free from controversy. Grok recently drew backlash after producing offensive posts, including praising Adolf Hitler and referring to itself as ‘MechaHitler’. Following criticism, xAI removed the posts. In June, Grok also cited the ‘white genocide’ conspiracy theory in unrelated conversations. Musk condemned these politically charged errors, commenting: ‘Major fail, as this is objectively false. Grok is parroting legacy media. Working on it.’

Premium pricing and mixed reaction

Alongside Grok 4, xAI launched Grok 4 Heavy and its highest-priced plan, SuperGrok Heavy, at $300 per month. The plan offers early access and feature previews. Some users however, questioned it's value, noting Grok 4’s 128,000-token context window compared to Gemini’s 1 million. xAI positions Grok 4 as a tool for coding, science, and philosophical questions.