Robots Atlas>ROBOTS ATLAS
Grok 4
AI Modelsโ€บGrok

Grok 4

4ย ยทย Family: Grok
xAI fourth-generation flagship model with native tool use (function calling) and real-time search integration. Available in standard and Heavy (SuperGrok Heavy) tiers. Released July 9, 2025.
โœ“ Activeโœ“ Public accessLLMMultimodalReasoning modelTool-using model๐Ÿ“ Grok
Context window
256K
tokens
Release date
9 July 2025
๐ŸขSpaceXAIProducer
Access:APIHostedDeployment:โ˜ Cloud

Overview

Grok 4 is xAI's flagship multimodal reasoning model released on 9 July 2025, alongside the Grok 4 Heavy variant available to SuperGrok Heavy subscribers. xAI claims Grok 4 outperforms rival models on benchmark tests (xAI does not publicly disclose specific numbers outside its promotional materials). The model features native tool use and built-in real-time search (Web Search and X Search). Knowledge cutoff: November 2024 (confirmed in docs.x.ai). Context window: 256K tokens. From 12 July 2025 it has been deployed in Tesla Model S, 3, X, Y, and Cybertruck via the 2025.26 software update. The same month xAI announced "Grok for Government" โ€” a contract worth up to $200M with the U.S. Department of Defense (alongside Anthropic, Google, and OpenAI). Since 10 August 2025, free users have had limited access (2 prompts every 2 hours).

Classification
LLMMultimodalReasoning modelTool-using model
Family: Grok
Access & deployment
APIHosted
Cloud
Weights: Closed
Key parameters
๐Ÿ“ Context: 256K
โœ“ Tools
๐Ÿ“ฅ Input: text, image

Technical specification

Context window
256K
tokens
Knowledge cutoff
1 Nov 2024
Knowledge boundary
License
Proprietary
Features:โœ“ Tool use
Modalities
โฌ‡ Input
textimage
โฌ† Output
text

Capabilities and applications

Native model capabilities
Reasoning
The model's ability to reason logically and solve complex problems.
Category: reasoning
Multi-step reasoning
Carrying out multi-step chains of reasoning across long, complex tasks.
Category: reasoning
Coding
Generating, analysing and modifying code in many programming languages. Covers writing functions, debugging, refactoring, code review, and creating tests. Measured by benchmarks such as HumanEval and SWE-bench.
Category: coding
Image understanding
Analysing and interpreting the content of images.
Category: vision
Multilingual
Competence in many natural languages (from a few to over a hundred): understanding, generation, translation, and code-switching within a single conversation. Frontier models support a wide range of languages with comparable quality.
Category: language
Long context
Support for large context windows โ€” tens to hundreds of thousands (or millions) of input tokens. Enables analysis of entire codebases, long documents, and many parallel conversations without losing earlier information. GPT-5.1 supports 400,000 tokens.
Category: language
Planning
Forming and executing action plans for complex tasks.
Category: planning
Function Calling
Category: planning
Structured output
Producing data in structured formats such as JSON.
Category: structured_generation

Technical architecture

Sources and related pages