Grok Build 0.1
xAI · Grok
xAI's lower-cost public-beta coding API model for fast agentic development, now separate from Grok Build's Grok 4.5 default.
Overview
Freshness note: Model capabilities, limits, and pricing can change quickly. This profile is a point-in-time snapshot last verified on July 10, 2026.
Grok Build 0.1 is xAI’s lower-cost coding model for agentic software work. xAI announced grok-build-0.1 for the xAI API in public beta on May 29, 2026, after listing it as early access in the May 19 release notes. It originally powered the Grok Build CLI, but Grok 4.5 became that product’s default model on July 8.
The practical positioning is now price- and compatibility-led: use this explicit model ID when you want the 256K coding route at $1 input / $2 output per million tokens or need the documented alias path from older Grok Code Fast configurations. Use Grok 4.5 when you want xAI’s newer flagship coding and agentic behavior.
Capabilities
Grok Build 0.1 is designed for agentic coding rather than plain code completion. xAI highlights web development, debugging, MCP support, function calling, structured outputs, and reasoning support. It also accepts text and image input and produces text output, which makes it useful for UI debugging, screenshot-guided code work, and agent harnesses that combine repository context with visual state.
xAI’s launch note says the model performs best in agentic harnesses such as Grok Build, Cursor, Hermes Agent, OpenClaw, Kilo Code, or OpenCode. That is the right mental model: the model matters, but the review loop, tools, file-editing policy, and test execution around it matter just as much.
Technical Details
Current xAI docs list:
- Model ID:
grok-build-0.1 - Aliases:
grok-code-fast-1,grok-code-fast,grok-code-fast-1-0825 - Context window: 256,000 tokens
- Modalities: text and image input, text output
- Capabilities: function calling, structured outputs, and reasoning
- Rate limits are model-, team-, and tier-dependent; xAI publishes current RPS and TPM limits in its live rate-limit table and console
xAI does not publish a separate max-output token cap on the model page at this snapshot. Here, maxOutput: 0 means “not separately published,” not a literal output limit.
Pricing & Access
Current xAI API pricing is:
- Input: $1.00 per 1M tokens
- Cached input: $0.20 per 1M tokens
- Output: $2.00 per 1M tokens
xAI also notes that requests exceeding the 200K-token context range can use different pricing. For large-context agent work, confirm the live pricing page and region-specific terms before making routing decisions.
Access is available through the xAI API in public beta. xAI’s launch note also says Grok Build 0.1 is available through OpenRouter and Vercel AI Gateway, but production planning should start from first-party xAI API terms.
Best Use Cases
Choose Grok Build 0.1 for scoped coding-agent tasks, repository edits, bug fixes, UI implementation, MCP-connected development workflows, and coding automation where speed and price matter. It is also the natural migration target for old grok-code-fast-1 configurations.
Avoid treating it as an autonomous senior engineer. It still needs narrow task framing, tests, diff review, and clear tool permissions. For harder coding or agentic knowledge work, compare it with Grok 4.5; for lower-cost general reasoning, compare it with Grok 4.3.
Comparisons
- Grok 4.5: Newer flagship and the current Grok Build default, with a 500K context window and higher token rates.
- Grok 4.3: Lower-cost 1M-context route for general reasoning and agentic assistant work.
- GPT-5.3-Codex: More mature inside OpenAI’s first-party Codex product surface.
- Claude Code / Claude Sonnet: Often stronger when the surrounding workflow depends on Anthropic’s coding-agent tooling and review conventions.