Grok 4.5 achieves an Artificial Analysis Intelligence Index score of 54, ranking fourth behind Fable 5, GPT-5.5, and Opus 4.8, positioning xAI back among North American frontier labs. The model's API pricing is $2/M for input and $6/M for output, offering roughly 2x token efficiency compared to leading models for comparable tasks. Grok 4.5 scores 83.3% on Terminal-Bench 2.1, nearly tying GPT-5.5, and 64.7% on SWE-Bench Pro, outperforming GPT-5.5.