Short answer: Grok Code Fast 1 was introduced as a tool-oriented model for agentic coding, including codebase questions, file edits, and bug fixes. Current xAI model-selection guidance recommends Grok 4.7 for code. For API use, avoid assuming that the legacy name grok-code-fast-1 still routes to the original model: xAI’s model page lists it as an alias for Grok Build 0.1, while a separate retirement notice describes redirects for the retired slugs it covers. The available details here do not establish whether that notice includes this alias, so confirm its routing and billing before deployment.
What Grok Code Fast 1 was built to do
In its August 2025 launch description, xAI positioned Grok Code Fast 1 for agentic coding and highlighted tools such as grep, terminal operations, and file editing. The description also named project scaffolding, codebase questions, and targeted bug fixes as intended tasks. These describe intended use, not guaranteed results for every codebase or development environment.
A September 2025 InfoQ report attributed a 70.8% SWE-Bench-Verified score to xAI’s internal evaluation suite and reported approximately 92 tokens per second and a 256,000-token context window. Treat these as dated, attributed figures—not a current, independently matched comparison or a prediction of performance on your repository. ()
Production API: model ID, price, and limits
The xAI model page for Grok Build 0.1 lists grok-code-fast-1, grok-code-fast, and grok-code-fast-1-0825 as aliases. It lists a 256,000-token context window and these API rates:
| Model listed on the page | Input | Cached input | Output |
|---|---|---|---|
| Grok Build 0.1 | $1.00 / 1M tokens | $0.20 / 1M tokens | $2.00 / 1M tokens |
A separate xAI retirement notice says requests to the retired model slugs covered by that notice redirect to Grok 4.3 after May 15, 2026. The available notice excerpt does not establish whether grok-code-fast-1 is among those slugs. So the alias listing and retirement notice alone do not settle how that name will route or which rate will apply. Use an explicit model ID and confirm routing and billing for your account.
For new coding integrations, xAI’s model-selection page recommends Grok 4.7. Its model page lists a 500,000-token context window and these rates:
| Model listed on the page | Input | Cached input | Output |
|---|---|---|---|
| Grok 4.7 | $2.00 / 1M tokens | $0.50 / 1M tokens | $6.00 / 1M tokens |
The page also lists rate limits of 150 requests per second and 50 million tokens per minute. The available pricing details do not establish whether these rates change at a particular prompt or context length. If your workload sends very long prompts, confirm the applicable pricing before estimating costs. The model-specific maximum output limit is not established by the available details; check current model documentation before relying on a particular output length.
Minimal API starting point
The xAI API reference identifies POST /v1/responses as an interface for text generation, reasoning, and tool use. Requests require input; specify the model with model. The API base URL is https://api.x.ai, and requests use a bearer API key. For example:
curl https://api.x.ai/v1/responses -H "Authorization: Bearer $XAI_API_KEY" -H "Content-Type: application/json" -d '{"model":"grok-4.7","input":"Review this function for bugs."}'
Create an xAI API key in the console. Confirm account access and billing eligibility before production use, and check current output and rate limits for your selected model and account. A successful API call does not, by itself, establish production suitability.
Availability is not the same as a current free offer
The August 2025 launch announcement described no-cost access for a limited time through selected partners, including GitHub Copilot, Cursor, Cline, Roo Code, Kilo Code, opencode, and Windsurf. That historical offer does not establish current partner availability or direct API pricing. Check the terms for the channel you plan to use; partner access, subscriptions, third-party providers, and direct API billing are separate routes.
Practical selection takeaway
- Considering the original coding workflow? Evaluate it on representative repository-search, edit, and terminal tasks. Launch-era descriptions and benchmark reporting are useful context, not a guarantee of results.
- Starting a new xAI API coding integration? Current model-selection guidance recommends Grok 4.7 for code. Use an explicit model ID and confirm its pricing, limits, and account access.
- Already sending
grok-code-fast-1? Do not assume it still identifies the original model or carries the original price. The available alias and retirement details do not establish the effective routing for that name; confirm behavior and billing before estimating cost or depending on model behavior.
In short, Grok Code Fast 1 is best understood as the launch-era name for xAI’s tool-oriented coding model. Its reported focus can inform a workflow evaluation, but for a new API integration, select and price the explicit model you intend to call.
Sources
Grok Code Fast 1 — xAI launch announcement
xAI Releases Grok Code Fast 1, a New Model for Agentic Coding — InfoQ
Grok Build 0.1 — xAI model documentation
May 15, 2026 Model Retirement — xAI
Grok 4.3 — xAI model documentation
Grok 4.7 — xAI model documentation