Grok 4.7 (xhigh) Demonstrates High Intelligence and Competitive Pricing, According to Artificial Analysis

According to an analysis by Artificial Analysis, the model scores 46 on the Intelligence Index, placing it well above the median score of 11 for comparable models in its price tier. During evaluation, the model generated 240M tokens, significantly more than the median of 92M tokens used by similar models.

In terms of pricing, Grok 4.7 (xhigh) is competitively priced at $0.00 per 1M input tokens and $0.00 per 1M output tokens, compared to a median of $0.17 and $0.57 respectively for similar models. The model is a proprietary reasoning model, meaning it uses extended thinking or chain-of-thought reasoning to process complex problems, though its specific model size and parameter count have not been disclosed by SpaceXAI.

PLUS ULTRAby Amenoyomi

The outstanding intelligence score of Grok 4.7 (xhigh) is rooted in its architectural approach as a reasoning model. Rather than providing an immediate response, the model employs "extended thinking" or "chain-of-thought" reasoning to systematically work through complex problems before delivering a final answer.

This internal reasoning process is directly reflected in the model's high output volume. During the Intelligence Index evaluation, Grok 4.7 (xhigh) generated 240 million tokens, which is more than 2.6 times the median of 92 million tokens used by comparable models in its price tier.