xAI introduces grok-code-fast-1, an artificial intelligence model designed to accelerate programming workflows, with support for multiple languages and tools, optimized performance, and cost-effective pricing for developers.
xAI has announced the launch of grok-code-fast-1, an artificial intelligence model developed specifically to optimize agile programming workflows and address the speed limitations faced by developers with current systems.
The new model was built from scratch with a completely new architecture, using a pre-training corpus rich in programming-related content. During post-training, the company employed high-quality datasets reflecting real-world pull requests and coding tasks.
A key feature of grok-code-fast-1 is its processing speed. According to xAI, the model can execute dozens of tools before the user finishes reading the first paragraph of the reasoning process. This speed results from innovative techniques developed by the company’s inference and supercomputing teams, along with caching optimizations that achieve hit rates above 90% when used with launch partners.
The model is versatile across the full software development stack, with particular proficiency in TypeScript, Python, Java, Rust, C++, and Go. It can complete common programming tasks with minimal supervision, from building projects from scratch to performing precise bug fixes and answering codebase queries.
In terms of performance, grok-code-fast-1 scored 70.8% on the full SWE-Bench-Verified set using xAI’s internal evaluation system. The company complements these metrics with routine human assessments conducted by experienced developers to measure end-to-end performance in everyday tasks.
The model is available with economical pricing: $0.20 per million input tokens, $1.50 per million output tokens, and $0.02 per million cached input tokens. For a limited time, xAI provides free access through launch partners including GitHub Copilot, Cursor, Cline, Roo Code, Kilo Code, opencode, and Windsurf.
xAI designed grok-code-fast-1 to be widely accessible, balancing performance and cost in a compact format effective for handling common programming tasks quickly and efficiently.
AI assistant developed by xAI that combines text processing and image generation. Integrated with X platform for real-time data access and DeepSearch function for advanced ...
11/05/2026
Thinking Machines Lab has published a research preview of TML-Interaction-Small, an interaction model designed to collaborate with the user in real ...
24/04/2026
DeepSeek releases a preview of its V4 family, two open-source models capable of processing up to one million tokens of context and competing with the ...
23/04/2026
OpenAI launches GPT-5.5, a model designed to handle complex tasks autonomously — coding, researching, analyzing data and operating a computer ...
21/04/2026
OpenAI introduces ChatGPT Images 2.0, an image generation model with greater precision, multilingual support, flexible aspect ratios and, for the ...