<ul data-eligibleForWebStory="false"><li>Elon Musk's xAI Holdings Corp. has launched a new reasoning-optimized large language model called Grok 4, designed for tasks like generating code.</li><li>Grok 4 was trained on the Colossus supercomputer with a capacity of up to 1 million graphic cards and can process prompts with up to 256,000 tokens. It offers enhanced reasoning capabilities compared to the previous version, Grok 3.</li><li>Grok 4 excelled in solving complex questions, achieving a score of over 44% on the Humanity's Last Exam benchmark dataset without external aid. It outperformed OpenAI's Deep Research tool.</li><li>xAI plans to introduce a subscription-based enhanced version, Grok 4 Heavy, priced at $300 per month. Future versions optimized for programming tasks, multimodal capabilities, and video generation are in the roadmap.</li></ul>

Elon Musk’s xAI sets AI benchmark records with new reasoning-optimized Grok 4 model

Discover more