Skip to content

ModelsxAI

xAI launches Grok 4 with native tool use and a Grok 4 Heavy multi-agent tier

xAI Grok 4 at a glance

Grok 4 was trained with reinforcement learning at pretraining scale to use tools such as code and search, launched in the API with a 256K context window, and came with a new SuperGrok Heavy tier.

What changed with xAI Grok 4

On July 9, 2025, xAI released Grok 4 to SuperGrok and Premium+ subscribers and through the xAI API. xAI said it used its 200,000-GPU Colossus cluster to run reinforcement learning at pretraining scale, teaching the model to reach for a code interpreter, web browsing and deep X search when needed.

xAI also introduced Grok 4 Heavy, which uses parallel test-time compute to consider several hypotheses at once, available through a new SuperGrok Heavy subscription. It reported 15.9% on ARC-AGI-2 for Grok 4, a new high for closed models at the time by its account, and the API offered a 256,000 token context window with text and vision input.

xAI Grok 4 in numbers

Values are copied from the linked source on the date shown. They are not IA Compa measurements.

Context windowas of · Source for Context window: x.ai (opens in a new tab)
256K tokens

How it compares

Models that were current when this was published (released in the six months before it, newest version of each line), so these figures are from that time. Under each chart is the current best. (Models)

Context window (tokens) · Grok 4higher is better

Current best (as of Nov 6, 2026): GPT-6 Cyber, 1.05M tokens

Where to compare xAI Grok 4

Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.

Why xAI Grok 4 matters for businesses

For businesses, Grok 4 added a strong tool-using model with real-time search across X and the web, which can help with market monitoring or social listening, but teams should validate its vendor-reported results and content policies against their own use cases.

Sources for this story

  1. Official
  2. Grok 4 model page (opens in a new tab)

    Artificial Analysis · artificialanalysis.ai

    Benchmark

Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.

  • #llm
  • #xai
  • #grok
  • #reasoning
  • #tool-use

Want this working in your business?

We turn new AI capabilities into agents, voice receptionists, and automations you can actually use.