xAI has introduced Grok Build 0.1, a coding-focused model available through its API in public beta. The company describes the model as built for agentic coding tasks, with a 256,000-token context window, text and image inputs, and pricing of $1 per million input tokens and $2 per million output tokens.
xAI has launched Grok Build 0.1, a coding-focused AI model now available through the xAI API in public beta.
In a company announcement titled “Grok Build 0.1 on API,” xAI describes grok-build-0.1 as its “latest and fastest coding model.” The company says the model is intended for agentic coding workflows, including web development, debugging, and Model Context Protocol support.
The xAI developer documentation similarly lists Grok Build 0.1 as a fast coding model trained for agentic coding. According to that documentation, the model supports both text and image input and offers a 256,000-token context window.
That long context window may be relevant for software-development use cases where a model needs to work across large codebases, documentation, logs, or multi-file project context. xAI’s own materials position the model around coding assistance rather than general-purpose chat.
According to xAI’s model documentation, Grok Build 0.1 is currently in early access and is priced at $1 per million input tokens and $2 per million output tokens. VKTR’s June 5 coverage also reported the public beta availability, the same $1/$2 per million token pricing, the 256,000-token context window, and generation speeds of more than 100 tokens per second.
The model is accessed via the xAI API under the model name grok-build-0.1, according to xAI’s announcement and developer documentation. The company’s framing suggests it is aimed at developers building coding agents or integrating automated coding support into software tools.
The release places xAI more directly into the market for coding-specific AI models, where developers compare systems on speed, cost, context length, tool use, and reliability on real repositories. xAI’s public materials emphasize fast generation, code-oriented training, and compatibility with agentic development patterns.
The company does not, in the cited materials, provide independent benchmark comparisons against rival coding models. As a result, claims about relative performance should be treated cautiously until third-party evaluations or reproducible benchmark results are available.
Still, the combination of public beta access, a large context window, multimodal input, and per-token pricing gives developers concrete parameters to test. For teams already using API-based coding assistants, Grok Build 0.1 may be evaluated on practical measures such as repository understanding, patch quality, debugging usefulness, latency, and cost under production workloads.
grok-build-0.1 and describes it as trained for agentic coding.Grok Build 0.1 is therefore best understood as an early-access coding model release rather than a fully proven category leader. The most meaningful evidence will come from developer testing and independent comparisons using real coding tasks.
xAI has launched Grok Build 0.1, a coding focused AI model now available through the xAI API in public beta.
The xAI developer documentation similarly lists Grok Build 0.1 as a fast coding model trained for agentic coding.
According to that documentation, the model supports both text and image input and offers a 256,000 token context window.
Continue reading