xAI, the AI startup owned by Elon Musk, has released Grok Code Fast 1, a new artificial intelligence model designed to help with coding. This lightweight AI tool comes with agentic capabilities, meaning it can perform tasks more independently and handle full-stack software development. It can be used to create both backend and frontend components of software, making it a versatile tool for developers.
The model is built on a new architecture, paired with programming-related content for pre-training and real-world coding examples for post-training. xAI claims that Grok Code Fast 1 solves common problems faced by other AI coding tools, which often slow down when executing complex functions or calling external tools. Grok Code Fast 1 is designed to generate output faster, making it efficient for coding workflows. It can also call tools like grep, terminal commands, and file editors during its tasks.
Grok Code Fast 1 supports multiple popular programming languages, including TypeScript, Python, Java, Rust, C++, and Go. According to xAI, the AI can complete common coding tasks with minimal supervision. Internal testing showed that the model scored 70.8 percent on the SWE-Bench-Verified benchmark, demonstrating strong performance for real-world coding challenges.
Starting August 28, xAI is offering Grok Code Fast 1 free of cost for seven days. During this period, developers can use the model on platforms like GitHub Copilot, Cursor, Cline, Kilo Code, Roo Code, opencode, and Windsurf. The company did not mention any limits on usage during this free period, giving developers a chance to explore the tool without paying.
After the free trial, Grok Code Fast 1 will be available on a pay-as-you-go model. The pricing will be $0.20 per million input tokens and $1.50 per million output tokens, with cached input tokens charged at $0.02 per million. This makes it a cost-effective option for developers looking to speed up coding tasks and build software efficiently.
xAI has highlighted that Grok Code Fast 1 is faster and more responsive than many rival coding tools, making it suitable for both professional developers and hobbyists. The model’s full-stack capabilities and agentic nature aim to simplify software development, allowing AI to handle more complex coding tasks independently.
This launch marks another step in integrating AI into software development, giving developers a powerful tool to save time and improve productivity.
