Elon Musk’s xAI launches Grok-Code-Fast-1, a free agentic coding model to rival OpenAI and Microsoft

Elon Musk’s artificial intelligence startup xAI is moving into the coding space. On Thursday, the company released a new agentic coding model called grok-code-fast-1, a tool it describes as “speedy and economical.” The release marks xAI’s push into one of the most competitive frontiers in AI: building assistants that can write and manage code.

The timing of the launch is hard to miss. Just three days earlier, xAI filed a lawsuit in U.S. federal court in Texas against Apple and OpenAI, accusing them of illegally conspiring to block competition in AI. Musk’s company is now taking that fight to the product level, squaring off against rivals that already dominate the space.

Introducing Grok Code Fast 1, a speedy and economical reasoning model that excels at agentic coding. Now available for free on GitHub Copilot, Cursor, Cline, Kilo Code, Roo Code, opencode, and Windsurf,” xAI said in a post on X.

Agentic coding tools are AI programs that can autonomously take on coding tasks, from writing snippets of code to debugging and suggesting improvements. It’s the type of product that’s quickly becoming a staple for developers.

xAI Enters AI Coding Race with Grok-Code-Fast-1: Musk’s Answer to GitHub Copilot and Codex

xAI says grok-code-fast-1 “delivers strong performance in an economical, compact form factor, making it a versatile choice for tackling common coding tasks quickly and cost-effectively.” For now, the model will be offered free for a limited time. GitHub Copilot and Windsurf are among the first partners to integrate it.

The move pits Musk’s startup directly against OpenAI and Microsoft, both of which have staked their bets on AI coding assistants. Microsoft rolled out a coding agent inside GitHub Copilot during its Build developer conference in May. CEO Satya Nadella has said that between 20% and 30% of all code across Microsoft products is now written by AI. OpenAI’s own coding agent, Codex, was made available to ChatGPT Plus users in June.

Inside Grok-Code-Fast-1

xAI said the new model was built from scratch on a lightweight architecture, paired with engineering tweaks aimed at boosting serving efficiency. The company described Grok-Code-Fast-1 as setting “a new standard for both speed and affordability,” positioning it as a leaner alternative to heavyweight rivals.

The model is available through the xAI API with pricing set at $0.20 per million input tokens, $1.50 per million output tokens, and $0.02 per million cached tokens. According to the company, it performs well across the full software stack but shows particular strength in TypeScript, Python, Java, Rust, C++, and Go.

xAI pointed to early examples of developers using the model, highlighting one who built a functional game (shown below) in a single day with Grok-Code-Fast-1. The startup said it emphasized human evaluations during training, focusing on how satisfied real developers felt using the tool, which it claimed helped the model earn ratings as fast, reliable, and economical for everyday coding tasks.

Grok Code Fast 1 is versatile across the full stack and is particularly strong at TypeScript, Python, Java, Rust, C++, and Go.

Using Grok Code Fast 1, @DannyLimanseta built the following game in a day. pic.twitter.com/rz2RgBno5l

— xAI (@xai) August 28, 2025

For its launch week, the model is being offered free on several agentic coding platforms, including Cursor, GitHub Copilot, Cline, Opencode, Windsurf, Roo Code, and Kilo Code. The company said it also released a guide with tips for developers to get the best results out of Grok-Code-Fast-1.

With grok-code-fast-1, Musk is betting that developers will opt for a leaner, faster option in a market where the giants already have a head start. For xAI, it’s both a strategic product release and a statement of intent: the company wants a seat at the table in the future of automated coding.