SpaceXAI Launches Advanced Grok 4.5 AI Model

SpaceXAI has announced the release of Grok 4.5, describing it as its most capable AI model to date. The new system is designed for software development, engineering, agentic workflows, and knowledge-intensive tasks, while also becoming the default model in Grok Build. It is already available through the company's developer platform and Cursor. Users in the European Union will gain access later.

The model was trained using large-scale infrastructure powered by tens of thousands of NVIDIA GB300 GPUs. Alongside extensive computing resources, the company emphasized improvements in dataset filtering, quality evaluation, and domain-specific data selection. These enhancements were designed to improve reasoning accuracy across programming, mathematics, science, and engineering.

According to SpaceXAI, Grok 4.5 delivers stronger performance on complex development tasks, including full application creation, advanced software engineering, and technical problem-solving. The company also highlighted expanded productivity features. The model can now generate sophisticated spreadsheets, prepare presentations with native design elements, create structured documents, and support long-running agentic workflows.

SpaceXAI says Grok 4.5 combines inference speeds of up to 80 tokens per second with roughly double the token efficiency of comparable leading AI models, which can help reduce execution time and operating costs. API pricing starts at $2 per million input tokens and $6 per million output tokens, positioning the model as a cost-effective option for developers and enterprise users.