Short answer
Yes. xAI describes grok-build-0.1 as a coding model for agentic tasks such as web development, debugging, and MCP support. If you searched for Grok Code Fast 1, note that xAI retired grok-code-fast-1 from its API effective May 15, 2026. The retirement page’s model-specific replacement table maps that slug to grok-build-0.1, while its general redirect text says retired slugs go to grok-4.3. Because those statements conflict, specify grok-build-0.1 directly for a new integration rather than relying on the old name’s routing. [1]
The names refer to related but distinct things: grok-code-fast-1 is the retired API model slug; grok-build-0.1 is the model name used in xAI’s API announcement; and Grok Build is xAI’s coding-agent product. xAI said in its May 2026 announcement that grok-build-0.1 was the model powering Grok Build at that time. [2]
What coding work is it meant for?
xAI says grok-build-0.1 was trained for agentic coding tasks including web development, debugging, and MCP support. Its model documentation lists a 256,000-token context window and capabilities including function calling and structured outputs. These are provider-described capabilities, not a guarantee that every task or repository will succeed. [2, 3]
The original Grok Code Fast 1 announcement described tool-oriented coding with grep, terminal operations, and file editing, and named TypeScript, Python, Java, Rust, C++, and Go. [5] A September 2025 InfoQ article reported a 70.8% SWE-Bench-Verified score using xAI’s internal evaluation suite. Treat that as a historical report about the earlier model—not an independently comparable result or verified performance measure for grok-build-0.1. [4]
API availability and pricing
xAI’s announcement says grok-build-0.1 is available through the xAI API in public beta. [2] Public-beta availability is not, by itself, a production-readiness recommendation: validate account access and limits, security and operational requirements, and performance on your workload before deployment.
A minimal text request uses the Responses API endpoint, bearer-token authentication, and an input string. The endpoint and request format below follow xAI’s official API quickstart; grok-build-0.1 is the model identifier listed in its model documentation. [3, 10]
curl https://api.x.ai/v1/responses \
-H "Authorization: Bearer $XAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "grok-build-0.1",
"input": "Find the bug in this function and suggest a fix: function median(a){a.sort();return a[a.length/2]}"
}'
The Responses API returns generated content in its output array. For a convenient text-only result, xAI’s quickstart shows the SDK’s response.output_text property. The response also includes usage information: record input and output token counts for cost tracking, and check usage.input_tokens_details.cached_tokens if you need to account for cached input. [10, 11, 12]
xAI’s pricing page lists these USD rates per 1 million tokens for grok-build-0.1: [6]
| Prompt context | Input | Cached input | Output |
|---|---|---|---|
| Short context | $1.00 | $0.20 | $2.00 |
| Long context (200,000 tokens or more) | $2.00 | $0.40 | $4.00 |
When a prompt reaches the model’s long-context threshold, the long-context rates apply to all tokens in that request. Requests sent to the US regional endpoint are billed at 1.1× the global token rates, a 10% premium. Check the current pricing documentation and your account’s applicable limits before deployment; this guide does not establish account-specific request or token limits. [6]
How to decide whether it fits your coding workload
It may be worth testing for workflows involving repository questions, focused bug fixes, or an agent that uses tools to inspect and edit code. xAI positions the model for agentic coding, but that positioning alone does not establish that it will be faster, less expensive, or more accurate than alternatives on your codebase. [2]
Run a small, controlled trial before committing:
- Choose representative tasks from your projects, such as a bug fix, a multi-file change, and a codebase question.
- Give each model the same instructions, repository context, tools, and acceptance tests.
- Track test results, review and rework time, completion time, and token cost.
- For agent workflows, check whether tool calls are appropriate and edits stay within scope.
- Compare the results with your quality, latency, and budget requirements—not just launch claims or a single benchmark.
For new API requests, specify grok-build-0.1 directly. The retirement page’s replacement table names it for grok-code-fast-1, but the same page’s general redirect statement names grok-4.3; avoid depending on the old slug when model choice and billing matter. [1]
What about Grok Build?
Grok Build is xAI’s coding-agent product. Its overview describes interactive and headless use, while xAI’s May 2026 API announcement said grok-build-0.1 was the model powering it at that time. The product and API model are related, but they are not interchangeable names. [2, 9]
Sources
Grok Code Fast 1 model documentation
xAI Releases Grok Code Fast 1, a New Model for Agentic Coding
Comparison with Chat Completions
Grok Build coding-agent overview