In a recent AI competition, Grok 4, a new artificial intelligence from xAI, outperformed the popular GPT-5. The contest took place in the ARC-AGI test, which measures how well AI can solve complex problems.
GPT-5 scored 65.7% on the first part of the test. But Grok 4 did much better, reportedly passing the test with ease. On the second part, GPT-5 only scored 9.9%, while Grok 4 solved the puzzles quickly and accurately.
Grok 4 also showed it is more cost-efficient. It uses less money and resources to complete each task compared to GPT-5, which costs about $0.51 per task.
Even GPT-5’s smaller versions, Mini and Nano, could not match Grok 4’s performance. Mini scored 54.3%, and Nano scored just 16.5%.
Fans of Grok 4 are celebrating online, sharing jokes and memes about how it defeated GPT-5. Many see Grok 4 as the new smartest AI around.
The future is uncertain, though. GPT-5 might come back stronger. For now, Grok 4 holds the top spot in this AI battle.