🚨 GROK 4 OUTTHINKS GPT 5 ON COMPLEX REASONING TEST In the latest ARC AGI 2 benchmark, Grok 4 scored about 16% beating GPT 5’s 9.9% but at a higher $2 to $4 per task compared to GPT 5’s $0.73. The test measures reasoning over memorization, and Grok 4 also led ARC AGI 1 with 68% to GPT 5’s 65.7%. GPT 5 still offers better value for now. Smaller model versions scored much lower, and early trials are underway for the interactive ARC AGI 3 puzzle test. This win boosts Grok 4 but the AI race is far from over. Source: The Decoder, @minchoi