Back to models
Chat Models

xAI

Grok 4.20 Non-Reasoning

Fastest

Non-reasoning Grok 4.20 for direct-response workflows, with 1M context, image input, and tool calling.

FastCost-efficient
Open AppBrowse All Models

At a glance

Know the model before you prompt.

xAI API specifications
Context window
1,000,000 tokens
Maximum output
Not verified
Inputs → output
Text + Images → Text
Knowledge cutoff
Not verified

API model ID: grok-4.20-0309-non-reasoning

Capabilities & boundaries

What it supports. Where the limits are.

Tool support requires the appropriate API integration; a supported tool is not automatically active in every chat.

Supported API features and tools

  • Text responses without the reasoning capability listed for the Reasoning variant
  • A 1,000,000-token context with text and image input
  • Function calling when tools are connected by an integration
  • Structured outputs for schema-driven extraction
  • Batch API support with a published 20% token discount

Before you choose

  • This variant is documented as non-reasoning. Do not offer reasoning-depth controls or Multi-Agent settings for it.
  • A large context is not a guarantee of reliable extraction from every part of a document; validate important fields against source evidence.
  • A maximum output ceiling and training cutoff are not verified in the reviewed card.
  • It returns text, not native image, audio or video output. Tool support does not mean browsing or execution is automatically enabled.

Non-reasoning model

The Markdown model card explicitly lists Reasoning: No. This is a separate non-reasoning model ID, not a low-effort setting on the Reasoning variant. Use clear constraints and verifiable output formats; do not assume an effort parameter adds reasoning to this model.

  • The exact ID is grok-4.20-0309-non-reasoning. The documented grok-4.20-non-reasoning and latest aliases point to this variant.
  • Some generic HTML capability labels can be ambiguous. The model's explicit Markdown capability flag identifies this variant as non-reasoning.
  • Validate function-call arguments and structured output. Neither a well-formed result nor the absence of reasoning proves that the answer is accurate.

Put it to work

Start with a more useful prompt.

Original examples from EZ Ai Assist. Adapt these to your task and the features available in your workspace.

Workflow 01

Classify support tickets

Use fixed labels and an explicit uncertainty option.

Classify each support ticket below using only these categories: billing, access, bug, feature request or unclear. Return ticket ID, category, a short evidence quote and whether human review is needed. Do not solve the tickets or invent details. If the evidence supports multiple categories, choose unclear and briefly explain the ambiguity.

Workflow 02

Extract documented fields

Keep missing information distinct from inferred values.

Extract the requested fields from these product records into a table: record ID, product name, stated version, stated availability and source section. Copy values faithfully, normalize only whitespace, and use not stated for missing information. Do not infer a version from a date or availability from marketing language. Flag contradictory values for review.

Workflow 03

Rewrite without changing claims

Make a short message clearer while preserving factual boundaries.

Rewrite the announcement below for clarity in 120 words or fewer. Preserve every factual claim, qualification, date and price; do not add benefits, urgency or promises. Keep the brand name exactly as written. After the rewrite, list any ambiguity that prevented a confident edit and confirm which details you preserved.

Developer reference

xAI API pricing

These are global xAI API reference prices, not EZ Ai Assist subscription prices.

View EZ Ai Assist plans
Standard processing · USD per 1,000,000 tokens
Token type< 200,000prompt tokens≥ 200,000prompt tokens
Input$1.25$2.50
Cached input$0.20$0.40
Output$2.50$5.00
Additional API fees · USD
UsageRate and unit
Web Search$5.00per 1,000 calls
X Search: posts$5.00per 1,000 posts fetched
X Search: profiles$10.00per 1,000 profiles fetched
Code execution$5.00per 1,000 calls
  • At 200,000 prompt tokens or more, doubled token rates apply to the entire request. These are the same published rates as the 4.20 Reasoning variant, not a lower per-token tier.
  • Batch API is supported with a 20% token discount. Priority processing is billed at 2× only when the response confirms that service tier.
  • Enabled server-side tools add usage and fees. X Search bills per fetched post/profile; Web Search and code execution use call-based fees.
  • Compare measured total request usage and answer quality rather than assuming non-reasoning is always the cheaper completed workflow. App billing is separate.

Common questions

A few things worth knowing.

What does Non-Reasoning mean here?

xAI's explicit model card marks this variant as not supporting reasoning. It uses its own grok-4.20-0309-non-reasoning ID. It is not simply the Reasoning model with a different prompt or a claimed low-effort setting.

What tasks should I evaluate it on?

Start with well-defined classification, field extraction or short rewrites where correctness can be checked directly. Include ambiguous and missing-data examples in your evaluation. For difficult multi-step work, compare against the Reasoning variant.

What context and output does it support?

The public card lists 1,000,000 tokens of context, text/image inputs and text output. A separate maximum output and knowledge cutoff were not verified. The app may impose smaller limits.

Can I enable reasoning with an effort control?

The model card does not establish such a control for this variant. Select and evaluate the separate Reasoning model if that is what your workflow needs. Multi-Agent settings also belong to a different model.

Can a non-reasoning model still call tools?

Yes. This card lists function calling and structured outputs despite the absence of reasoning. Tools must still be configured, authorized and checked by the integration; they are not automatically connected in a chat.

Is its per-token price lower than Reasoning?

No. Both cards list $1.25 input, $0.20 cached input and $2.50 output per million below 200,000 prompt tokens, with doubled rates at or above the threshold. Total usage can differ, so compare complete task costs.

Does Batch support or API pricing determine my app plan?

No. xAI lists a 20% Batch token discount, but that is a developer API option, not an EZ Ai Assist subscription promise. Check the app for current access, controls, tools and pricing.

Check the source

Official documentation

Specifications and API prices checked on . Example prompts and workflow advice are editorial guidance from EZ Ai Assist.

Same provider

More from xAI

xAI

Grok 4.7

Coding and knowledge-work model with 500K context, image input, and configurable reasoning effort.

xAI

Grok 4.6

Coding and agentic model with 500K context, structured outputs, and four reasoning effort levels.

xAI

Grok 4.5

Coding and engineering model with 500K context, image input, tool calling, and configurable reasoning.

xAI

Grok 4.3

1M-context model for instruction-driven work and tool calling, with configurable reasoning that can be disabled.