A New Foundation for Enterprise AI
SpaceXAI has officially launched its latest artificial intelligence model, Grok 4.5, marking a significant milestone in the company’s strategy to capture the enterprise coding and productivity market. Announced by CEO Elon Musk, the model is built on the new V9 foundation, boasting 1.5 trillion parameters. The release follows a period of private beta testing across SpaceX and Tesla, positioning the model as a direct competitor to Anthropic’s Opus-class systems.
Strategic Integration and Performance
A core component of this release is the deep integration of Cursor’s AI coding agent technology, acquired by SpaceXAI earlier this year. According to the company, Grok 4.5 is specifically engineered for complex software engineering, science, and mathematical tasks. The model is trained on tens of thousands of Nvidia GB300 GPUs, utilizing advanced reinforcement learning to improve reasoning across multi-step technical projects.
Company data indicates that Grok 4.5 operates at speeds of 80 tokens per second (TPS). SpaceXAI claims the model achieves approximately twice the token efficiency of comparable industry leaders, allowing it to solve complex engineering tasks in significantly fewer steps. This efficiency is reflected in the pricing model, set at $2 per million input tokens and $6 per million output tokens.
Market Context
The launch arrives during a period of intense activity within the AI sector. OpenAI recently announced its GPT-5.6 ‘Sol’ model, while Anthropic continues to refine its Mythos and Sonnet 5 offerings. For SpaceXAI, Grok 4.5 serves as a critical revenue engine, with the company aiming to monetize its infrastructure through both API access and integrated enterprise tools. While currently available in most regions, the company noted that EU availability is expected to follow in mid-July.