xAI
Grok 4.7
grok-4.7
Total Context
500K
Max Output
500K
Released
N/A
Input$1.6 / M tokens
Output$4.8 / M tokens
Grok 4.20
grok-4.20Grok-4.20 is presented as a high-speed xAI reasoning model with strong agentic tool calling, strict prompt adherence, low hallucination, and a very large context window in third-party cards. It is meant for long-form analysis, deep research, and multi-step agent workflows. The strongest description should emphasize speed plus long-context reasoning, not just raw conversational personality.
Context Window
1M tokens
Maximum Output
30K tokens
Release Date
Mar 9, 2026
Modalities
Grok 4.20 Pricing
| Input Price | Output Price | Cache Read |
|---|---|---|
| $1.25/M | $2.5/M | $0.2/M |
Grok 4.20 API Capabilities
Reasoning
Supported
Tool calling
Supported
Temperature parameter
Supported
Attachments
Supported
Knowledge Base
—
Endpoint Protocols
Completions APIMessages APIgemini
Grok 4.20 Model Highlights
Grok 4.20 combines reasoning, strict instruction following, image understanding and million-token context for demanding agentic and analytical workloads.
Deliberate Reasoning
The reasoning variant can think before responding, supporting tasks that require multi-step analysis and careful evaluation.
Strict Instruction Following
The model is designed to follow detailed prompts consistently, helping applications enforce task requirements and output constraints.
Multimodal Long Context
Text and image inputs can be analyzed within a 1,000,000-token context window, accommodating extensive source material in one task.
Grok 4.20 Use Cases
Grok 4.20 is suited to complex investigation, long-form technical analysis and multimodal agent workflows requiring precise adherence to instructions.
Complex Technical Investigation
Work through multi-stage engineering or scientific questions, evaluate competing explanations and produce a reasoned conclusion.
Long-Form Document Review
Analyze large document collections against detailed criteria, trace relevant evidence and return a structured assessment.
Multimodal Agent Workflows
Interpret text and visual inputs, reason about the requested objective and carry out controlled multi-step tasks with external systems.
How to Use Grok 4.20 via the TokenHub API
import OpenAI from "openai"
const client = new OpenAI({
apiKey: process.env.TOKENHUB_API_KEY,
baseURL: "https://us-api.tokenhub.com/v1",
})
const result = await client.chat.completions.create({})
console.log(result.choices[0]?.message?.content)Grok 4.20 Benchmarks
Grok 4.20 0309 (Reasoning)
| Index score | ||
|---|---|---|
| Artificial Analysis Intelligence Index | Artificial Analysis broad capability aggregate | 36.5 |
| Artificial Analysis Coding Index | Artificial Analysis software task aggregate | 42.2 |
| Knowledge & Reasoning | ||
| GPQA | Advanced science problem solving | 88.5% |
| HLE | Broad expert-level exam set | 30% |
| Coding & Engineering | ||
| SciCode | Scientific coding challenges | 44.7% |
| Terminal-Bench Hard | Hard terminal task execution | 40.9% |
| Instruction Following & Agent Tasks | ||
| IFBench | Prompt constraint adherence | 82.9% |
| AA-LCR | Long-context reasoning | 59% |
| τ²-Bench | Agent workflow tasks | 96.5% |
Metrics sourced from Artificial Analysis
Frequently asked questions about Grok 4.20
Understand what Grok 4.20 is, its best uses, distinguishing strengths, practical tradeoffs, and safe TokenHub integration guidance.
Where does Grok 4.20 sit within its provider’s model family?+
Grok 4.20 is xAI’s high-performance Grok 4.20 model for reasoning, long context, and agentic tool calling. It is a beta model, so validate latency, output consistency, and supported features before production use.
Which production scenarios suit Grok 4.20?+
Best-fit scenarios include reliable execution of multi-step agent workflows, analysis of long documents and datasets, and complex multi-step reasoning. Test representative inputs and define measurable acceptance criteria before production.
What makes Grok 4.20 stand out for analysis of long documents and datasets?+
Key strengths include effective use of tools and function calls, strict adherence to prompts, and configurable reasoning effort. This combination is especially useful for analysis of long documents and datasets.
What tradeoffs should developers consider with Grok 4.20?+
Consider another model when the application needs production behavior that is already fully stabilized, the task is simple enough for a non-reasoning variant, or the workflow cannot include human review for important decisions. Verify important factual, legal, financial, medical, or operational outputs with qualified human review.
How can a team safely start using Grok 4.20 on TokenHub?+
In TokenHub, select the exact model identifier displayed for Grok 4.20, use the endpoint documented for your account, and authenticate with your TokenHub credentials. Check the TokenHub page for the exact Grok identifier, available reasoning controls, tool access, supported inputs, and current beta or release status.
Ready to use Grok 4.20?
Use one API key to access Grok 4.20 and more AI models through TokenHub.
Media and Discussions
Selected public videos and posts related to this model.
X (Twitter)
Reddit
YouTube