Elon Musk's artificial intelligence (AI) company, SpaceXAI, has unveiled its new AI model, 'Grok 4.5.' The model focuses on coding and task automation, aiming at the enterprise AI tools market.
According to Reuters on July 8, SpaceXAI introduced Grok 4.5 as "the most intelligent AI system" in its lineup.
Grok 4.5 is optimized for automation tasks that involve coding, knowledge work, and executing user instructions in multiple stages.
This model was trained in collaboration with the AI coding tool Cursor. Last month, SpaceXAI announced plans to acquire Cursor's developer, Anisphere, in a stock exchange deal valued at $60 billion.
Performance-wise, Grok 4.5 has reached levels comparable to leading models in certain coding metrics.
According to data released by SpaceXAI, Grok 4.5 scored 83.3% on the Terminal-Bench 2.1, which evaluates coding capabilities based on command input. This score is similar to GPT-5.5's 83.4% and higher than Claude Opus 4.8's 78.9%.
However, it did not outperform all competing models in every evaluation. In the SWE-Bench Pro, which assesses actual software problem-solving abilities, Grok 4.5 scored lower than Claude Opus 4.8 but higher than GPT-5.5.
The pricing for Grok 4.5 is set lower than its competitors. The usage fee is $2 per million input tokens and $6 per million output tokens. Tokens are the basic units used by AI to process sentences. In comparison, Claude Opus 4.8 charges $5 for input and $25 for output.
SpaceXAI stated, "Grok 4.5 can be utilized not only for coding but also for creating Excel models, PowerPoint charts, and Word documents."
This model will be available through Grok Build, Cursor, and developer service screens, starting mid-July in the European Union.
Musk noted on X (formerly Twitter) that "Grok 4.5 is an Opus-level model but faster, more token-efficient, and lower in cost."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.