Grok 4.20 (Non-Reasoning)

Grok 4.20 Non-Reasoning is an xAI generative model designed to deliver strong general-purpose language and vision capabilities without explicit chain-of-thought reasoning at inference time. The model carries the identifier grok-4.20-0309-non-reasoning and is positioned alongside a sibling Grok 4.20 Reasoning variant within xAI's broader Grok lineup, suggesting a deliberate split between fast, non-reasoning responses and a more deliberate reasoning mode. Its intended use centers on conversational assistants, retrieval-augmented workflows, and developer-facing applications that benefit from tool integration and structured outputs.

In practical terms, the model accepts text and image inputs and produces text outputs, with tool calling, code execution, web search, streaming, batch processing, prompt caching, and structured output all supported. Pricing on the reference page is listed at $1.25 per million input tokens, $2.50 per million output tokens, and $0.20 per million cached input tokens, with a 20% discount available for asynchronous batch requests. The 200K context tier carries its own higher rate band, making the model a reasonable fit for moderate-length document analysis and agentic pipelines that need reliable tool use without the latency overhead of explicit reasoning steps.