Grok

Overview

This is a model family overview. For version-specific details, see the individual model entries linked below.

Grok is xAI’s model family for reasoning-heavy assistants, tool-using agents, coding routes, and the consumer Grok product. The May 2026 lineup is less segmented than it was in early spring: xAI now strongly recommends Grok 4.3 for general API callers outside specialized audio, image, and video endpoints, while several older Grok text, coding, and image slugs moved into post-retirement redirect behavior on May 15, 2026 at 12:00 PM PT.

Current Latest

Grok 4.3 is xAI’s current recommended primary API caller. It rolled out across xAI’s surfaces in mid-to-late April 2026, with public pricing of $1.25 per million input tokens and $2.50 per million output tokens, and a 1M-token context window. xAI’s developer guidance is to default to Grok 4.3 for general workloads outside specialized audio, image, and video endpoints.

The most relevant Grok entries at this snapshot:

• Grok 4.3 as the recommended default for reasoning, non-reasoning, coding, agentic, and long-context API work.
• Grok 4.20 as a 2M-token preview/enterprise research lane for teams explicitly testing that beta family.
• Grok Imagine for image and video generation through xAI’s media endpoints.
• Grok 4, Grok 4.1 Fast, and Grok Code Fast 1 as deprecated snapshots whose old slugs now redirect to Grok 4.3 with documented reasoning-effort defaults.

Strengths

The family is strongest when a team wants xAI’s combination of reasoning, tool calling, and live-information access. xAI’s current docs also make it clear that Grok is designed to work with web search, X search, code execution, collections, connectors, and MCP-style tool flows rather than remaining a plain text model.

Pricing segmentation is also clearer than before. Grok 4.3 sits at the recommended-default tier and now absorbs the migration path for retired Grok 4, 4.1 Fast, Grok 3, and Code Fast routes. Requests to deprecated text slugs are billed at Grok 4.3 rates after the redirect, so explicit model selection is still safer than leaving old aliases in production.

When to Choose Grok

Choose Grok when the workflow depends on tool calls, current information, or xAI-native product alignment. It is especially relevant for search-heavy assistants, long-context agents, coding tools, and teams already evaluating Grok as both a consumer assistant and an API stack. For new text or coding work, start with Grok 4.3 rather than relying on deprecated redirect slugs.

For teams that need slower-moving documentation, stronger governance defaults, or less product churn, Anthropic and OpenAI remain easier operational baselines. Grok is improving quickly, but it still benefits from closer monitoring than the most mature alternatives.

Access

• xAI API
• Grok web and mobile apps
• X ecosystem consumer surfaces
• xAI enterprise and beta access paths for newer Grok releases