xAI retired Grok Code Fast 1 on May 15, 2026, consolidating its agentic coding offerings under Grok Build 0.1. The older model, which launched in August 2025 as a low-cost option for terminal and IDE workflows, is no longer available as a distinct product. Developers using the legacy API aliases will now automatically receive Grok Build 0.1 instead, though xAI recommends explicitly migrating to avoid confusion about model behavior and pricing.

What Was Grok Code Fast 1 and Why Did It Matter?

Grok Code Fast 1 was designed as a responsive, lightweight coding agent trained for iterative development work. Unlike traditional code-completion tools that generate snippets in isolation, this model was built to work inside an agent harness, meaning it could reason through multiple steps, use terminal commands, edit files, and search repositories in sequence. The model supported a broad range of programming languages, including TypeScript, Python, Java, Rust, C++, and Go.

At launch, the pricing reflected its positioning as a budget-friendly option. Input tokens cost $0.20 per million, cached input tokens cost $0.02 per million, and output tokens cost $1.50 per million. Those rates are now obsolete. The model was available through multiple partners, including GitHub Copilot, Cursor, Cline, Roo Code, Kilo Code, OpenCode, and Windsurf, giving it broad reach across the developer ecosystem.

How Should Developers Migrate From the Old Model?

• Explicit Migration: xAI recommends that teams explicitly update their configurations to use Grok Build 0.1 rather than relying on automatic routing of old API aliases, which ensures clarity about which model is running and what pricing applies.
• Review Old Invoices: Teams should audit past invoices and configurations tied to Grok Code Fast 1 to understand historical costs and identify any systems still pointing to the retired model's slugs.
• Evaluate Successor Capabilities: Grok Build 0.1 prioritizes speed and low latency over broader reasoning, so teams needing stronger analytical reasoning may want to consider Grok 4.5 as an alternative for certain workloads.

What Are the Alternatives for Developers?

xAI's consolidation leaves developers with several options depending on their priorities. For agentic coding and web-development workflows, Grok Build 0.1 is the recommended successor, maintaining the speed-first approach of its predecessor. However, teams that value deeper reasoning and broader agentic capabilities may find Grok 4.5 more suitable, even if it sacrifices some latency.

Beyond xAI's ecosystem, developers can consider Claude Code for mature terminal-based agent workflows powered by Anthropic's current coding models, or OpenAI Codex for repository-aware coding agents with parallel task capabilities and code review features. The retirement of Grok Code Fast 1 reflects a broader industry trend toward consolidating multiple models into fewer, more specialized offerings rather than maintaining a wide portfolio of overlapping tools.

Why Did xAI Retire the Model?

While xAI has not publicly detailed the reasoning behind the retirement, consolidating Grok Code Fast 1 into Grok Build 0.1 simplifies the company's product line and allows it to focus engineering resources on a single agentic coding model. The move also eliminates the need to maintain separate pricing, documentation, and support channels for two overlapping products. Developers who were using Grok Code Fast 1 through launch partners will continue to receive service, but under the new model's terms and pricing structure.

The transition underscores a key challenge in the rapidly evolving AI coding space: maintaining backward compatibility while advancing the underlying technology. By routing old API calls to the new model, xAI avoided a hard cutoff that would have broken existing integrations, but the company is encouraging teams to migrate explicitly to avoid surprises about model behavior or cost changes.