Grok 4 just broke away from the pack on the ARC-AGI-2 leaderboard — an independent third-party assessment.
• Elon Musk: on academic tests, xAI’s Grok 4 is better than almost all PhDs in all fields (!)
• Over 2x better than the best from OpenAI… and rapidly improving

Webcast introducing Grok 4, the world’s most powerful AI model.

Summary of benchmarks

2 responses to “Grok 4 and the Exponential Approach to Human Intelligence”

  1. sunrizr Avatar
    sunrizr

    Good. perhaps a chance at progress away from establishment script:)

    Reply
  2. sbove Avatar
    sbove

    Grok 3 was already like having an infinitely patient and hard working top ranked PhD professor in every subject on call 24/7. Astounding progress in training/model building, reasoning and textual-linguistic logic, not to mention the new voice and image/video synthesis.

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *