xAI debuts Grok 4.5 with faster coding, agentic tasks, advanced reasoning, and lower costs

xAI has launched Grok 4.5, its latest generative AI model, now available on X, the web, and iOS and Android apps. According to the company, Grok 4.5 is its smartest model so far, built for complex coding, agentic tasks, and advanced knowledge work, with training influenced by Cursor. The model was trained on data that includes coding, science, engineering, and mathematics, enabling it to handle real engineering tasks and outperform similar leading AI models.

Following these improvements, Grok 4.5 offers strong abilities in software development, handling challenging Rust and C/C++ problems, and delivering end-to-end app builds directly from user prompts. With a serving speed of 80 tokens per second and doubled token efficiency compared to other top models, it delivers faster responses while keeping costs low.

Within Grok Build, Grok 4.5 is the default model, supporting sophisticated Excel workflows such as researching and generating multi-sheet spreadsheets, employing complex formulas, and including notes for future reference. Subsequent enhancements extend to PowerPoint and Word, where the model constructs native diagrams, drafts slide content, and writes clear prose for documents.

Grok 4.5 is priced at $2 per million input tokens and $6 per million output tokens. It is available in Grok Build, on all Cursor plans, and through the SpaceXAI console, with EU support planned for mid-July.