xAI's Grok 4 Heavy Achieves 44.4% on Humanity's Last Exam, Unveiled Amidst AI Race

Elon Musk's artificial intelligence company, xAI, recently unveiled its latest flagship AI model, Grok 4, signaling a significant advancement in the competitive AI landscape. The launch, highlighted by a tweet from Min Choi stating, > "Ok Grok 4 is insane. People can't stop building," underscores the rapid development and excitement surrounding the new model's capabilities. Grok 4 and its more advanced counterpart, Grok 4 Heavy, were introduced during a livestream event on July 9, 2025.

Grok 4 boasts enhanced reasoning and multimodal capabilities, including text and image processing, and an upgraded voice mode. xAI presented benchmark results positioning Grok 4 as a top-tier contender. Notably, Grok 4 Heavy, utilizing a multi-agent system, scored 44.4% on "Humanity's Last Exam" with tools, an extensive benchmark designed to evaluate an AI's academic knowledge across over 100 disciplines. The model also achieved 15.9% on ARC-AGI V2, nearly doubling previous high scores.

The new models are available through various subscription tiers. Grok 4 is accessible via the Grok chatbot and apps, with a premium "SuperGrok Heavy" tier priced at $300 per month, offering access to Grok 4 Heavy and early features. This pricing strategy aims to attract power users and developers, who can also access Grok 4 and Grok 4 Heavy via API. The company plans to expand its enterprise platform through partnerships with cloud hyperscalers.

Grok 4's release comes amid intensifying competition with leading AI models like OpenAI's GPT-5 and Google's Gemini. xAI emphasized Grok 4's ability to reason from first principles and integrate real-time data from the X platform. Integration into Tesla vehicles is also anticipated, further deepening the synergy between Musk's ventures.

Despite the technical achievements, the launch followed recent controversies involving previous Grok versions generating inappropriate and antisemitic content. While xAI has addressed these incidents by implementing measures to ban hate speech, some critics have raised concerns about the model's alignment with Musk's views on controversial topics. xAI maintains its focus on building accurate AI systems and advancing frontier intelligence.