Short answer
Grok Code Fast 1 was designed for agentic coding, but xAI’s migration documentation says the grok-code-fast-1 slug was retired from the API effective May 15, 2026. For a new integration, use the explicit model name grok-build-0.1 and evaluate it on your own coding tasks. The API announcement described Grok Build 0.1 as available in public beta; that is API access, not a production-service guarantee. (xAI model documentation, Grok Build 0.1 announcement)
There is a relevant naming and billing wrinkle: the model page lists grok-code-fast-1 as an alias for Grok Build 0.1, while the migration guide says requests to retired slugs are redirected to Grok 4.3 and billed at Grok 4.3 rates. Because those statements do not give a consistent billing outcome for the old slug, specify grok-build-0.1 directly rather than relying on the alias or redirect. (model page, migration guide)
What Grok Code Fast 1 was designed for
In its 2025 announcement, xAI described Grok Code Fast 1 as a model for agentic coding tasks, including work with coding tools and codebases. InfoQ’s report on the launch named project scaffolding, codebase questions, and bug fixes as intended tasks, and listed TypeScript, Python, Java, Rust, C++, and Go as language examples. These descriptions explain the model’s intended fit; they do not guarantee successful task completion or equal performance across languages. (xAI announcement, InfoQ report)
It may be worth evaluating for an interactive coding agent or tool-using development workflow. The original launch report’s benchmark and speed figures are historical, attributed claims—not current deployment guarantees or a controlled comparison against alternatives. Test the current API model and your own agent setup before making a selection.
API access and capabilities
xAI’s current model page identifies Grok Build 0.1 as the model name grok-build-0.1. It lists a 256,000-token context window, function calling, and structured outputs. The listed regions are us-east-1 and us-west-2; listed rate limits are 37 requests per second and 10,000,000 tokens per minute. Check the model page and your account’s applicable limits before planning traffic. (xAI model documentation)
xAI’s API announcement provides this minimal request using the Responses endpoint and bearer authentication. Set XAI_API_KEY to a valid xAI API key before running it:
curl https://api.x.ai/v1/responses \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $XAI_API_KEY" \
-d '{
"model": "grok-build-0.1",
"input": "Find and fix the bug in this function: ..."
}'
This sends a coding prompt and requests a model response; it does not, by itself, create a coding agent. Your application must provide any tools, decide whether tool calls should execute, and validate proposed changes. xAI’s announcement describes the API as being in public beta, so assess whether that status and the terms available to you fit your intended use. (xAI API announcement)
API pricing and a cost example
The xAI model page lists these USD rates per 1 million tokens for Grok Build 0.1:
- Input: $1.00
- Cached input: $0.20
- Output: $2.00
The page also says requests exceeding the 200,000-token context window are charged at different rates, but the retrieved pricing information does not specify those higher-context rates. Check the applicable price before estimating a request that exceeds that threshold. (xAI model documentation)
Example: For one request with 100,000 uncached input tokens and 50,000 output tokens, the listed standard rates give an estimate of $0.20: $0.10 for input plus $0.10 for output. This example assumes the request is not subject to higher-context pricing and excludes any other charges.
Don’t rely on the retired slug’s billing
The migration guide says requests using retired slugs are redirected to Grok 4.3 and billed at Grok 4.3 rates: $1.25 per million input tokens and $2.50 per million output tokens. Its replacement table maps grok-code-fast-1 to grok-build-0.1, while the Grok Build 0.1 model page also lists the old name as an alias. Because the migration guide’s redirect and the model page’s alias listing leave the old slug’s billing behavior unclear, use grok-build-0.1 explicitly and confirm the rate shown for your API use. (xAI migration guide, xAI model page)
The original no-cost launch-partner offer was described as limited-time in 2025. It does not establish current API pricing or current availability through coding products. (xAI’s original announcement)
Grok Build 0.1 or Grok 4.7?
xAI’s current model catalog recommends Grok 4.7 for code and describes it as its flagship model for code and other uses. Its listed context window is 500,000 tokens, with API rates of $2 per million input tokens and $6 per million output tokens. Grok Build 0.1 is listed at lower standard per-token rates and has a coding-focused description. These specifications can help narrow the options, but they do not establish which model will perform better on your codebase. (xAI model catalog)
- Evaluate Grok Build 0.1 if its coding focus, listed features, limits, and public-beta terms fit your agent workflow.
- Consider Grok 4.7 if you want the model xAI currently recommends for code, or if its larger listed context window suits your workload; account for its higher listed token rates.
- Compare them on representative work if the decision is consequential. Use the same codebase tasks, tools, and API route, then measure test results, human correction needed, latency, and actual token cost. This is a suggested evaluation method, not a test performed for this guide.
Bottom line
Grok Code Fast 1’s agentic-coding focus helps explain the model’s original purpose, but the API choice now requires attention to its migration status. For a new request, specify grok-build-0.1, not the retired slug. Check the listed rates and limits, confirm the higher-context price if relevant, and decide whether public-beta access suits your use case. Then compare it with Grok 4.7 on the coding tasks and cost assumptions that matter to your application.
Sources
Grok Code Fast 1 / Grok Build 0.1 — xAI API model documentation
May 15, 2026 Model Retirement — xAI API documentation
Grok Models & Pricing — xAI API documentation
Grok Code Fast 1 — xAI announcement
xAI Releases Grok Code Fast 1, a New Model for Agentic Coding — InfoQ