Elon Musk’s AI-powered company, xAI, launched its frontier Grok 4 AI models with groundbreaking benchmark numbers, aiming to outdo major players such as ChatGPT and Gemini.
xAI CEO Elon Musk announced the flagship Grok 4 models, including Grok 4 and Grok 4 Heavy, during a livestream event.
The recently launched xAI reasoning models consist of a comprehensive suite of cutting-edge features, including image analysis and complex question answering.
To note, in the recently introduced ARC-AGI-2 benchmark, Grok 4 accomplished an impressive 15.9%, marking the highest score to date.
The model scored double that of Claude Opus 4 and OpenAI o3, making it the frontier AI model among all the flagship AI models launched by any AI lab.
In recent months, Grok has been rapidly incorporated into the social platform now owned by xAI, called X (formerly Twitter).
During the livestream, the tech billionaire claimed that “With respect to academic questions, Grok 4 is better than PhD level in every subject, no exceptions,” he said.
However, he admitted the model occasionally lacks common sense and will evolve more to achieve perfection.
xAI Grok 4 pricing
xAI revealed a subscription plan called SuperGrok Heavy, which costs $300 per month and provides access to the Grok 4 Heavy model.
