Features
Models
The core Grok language models from xAI.
Grok is the name of xAI's family of large language models. When people say they are using Grok, they usually mean they are interacting with one of these models through chat, an API, or another product surface such as Grok Bot, Grok Build, Imagine, or Voice.
Models are the underlying intelligence layer. They generate text, reason over problems, write and edit code, and support multi-step workflows when paired with tools. Everything else in the Grok product line is built on top of that layer.
Recent public releases
Public coverage on Grokati focuses on releases xAI has announced openly. That includes Grok 4.5, rolled out more widely after a beta period, and Grok 4.6, positioned for long-running agents, coding, and more ambitious interactive work.
Official Grok 4.6 materials described improvements for sustained multi-step tasks and listed availability across Grok Build, Cursor, the API, and selected partners.
What models are used for
- General chat and reasoning — answering questions, analysing information, drafting and editing text
- Coding and technical work — writing, reviewing, and working across codebases
- Agentic workflows — multi-step tasks when the model is connected to tools or harnesses
Models vs other Grok products
- Grok Build — harness for coding and long-running agentic work
- Grok Bot — teammate-style agents that use tools and return finished work
- Imagine — image and video generation and editing
- Voice — spoken interaction, including connectors in voice mode
The model is the engine. Those products are different ways of driving it.
How Grokati covers models
Under the Models category you will find launches, limit changes, and official statements about model behaviour. We do not invent benchmarks or roadmap items.