Trained on large volumes of real developer-tool interaction; available and aggressively priced.
Grok 4.5 is the model that put xAI on the independent map: an Artificial Analysis intelligence index of 54 at just $2 / $6 per million tokens, one of the better price-to-capability ratios in the balanced tier. It is available now, aggressively priced, and trained on large volumes of real developer-tool interaction.
Two limits shape where it fits. A context surcharge above 200K tokens roughly doubles rates, and the ~500K window trails the 1M-class leaders on very long inputs. It has also been superseded at the top of xAI's line by Grok 4.6 — though for now 4.5 is the xAI model with published independent numbers, which its newer sibling still lacks.
For coding and general work at a low, predictable price, provided inputs stay under the surcharge threshold, Grok 4.5 is a credible pick and currently the better-documented of xAI's two frontier-adjacent models.
FIELD NOTE — Context surcharge above 200K roughly doubles rates.
VISIT OFFICIAL SITE →