AI Rivalry Intensifies as Grok Emerges Challenging OpenAI’s New Model

Musk initiated the discussion shortly after GPT-5 became available, initially stating that Grok 4 Heavy was “smarter two weeks ago than GPT5 is now.” He further claimed Grok 4 achieved superior performance on the Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI), a benchmark designed to assess general AI intelligence by testing reasoning abilities, problem-solving skills, and adaptability. The ARC Prize website describes it as measuring an AI’s capacity to perform novel tasks with minimal prior training.