← Models

Grok 4

xAI · Grok

xAI's flagship Grok model focused on frontier reasoning, long context, and tool-oriented assistant workflows.

Part of Grok family · Other versions: Grok 4.20
Type
language
Context
256K tokens
Max Output
256K tokens
Status
current
Input
$3/1M tok
Output
$15/1M tok
API Access
Yes
License
proprietary
reasoning tool-use agentic coding long-context real-time-data
Released July 2025 · Updated March 6, 2026

Overview

Freshness note: Model capabilities, limits, and pricing can change quickly. This profile is a point-in-time snapshot last verified on February 15, 2026.

Grok 4 is the baseline Grok 4 generation model for advanced language and agent workflows. Since its July 9, 2025 launch, xAI has continued the line with Grok 4.1 updates for product surfaces, but Grok 4 remains the clearest published technical reference point in xAI’s official model announcements.

In practical product design, Grok 4 is the premium tier for difficult queries, with faster or cheaper variants used for high-volume traffic.

Capabilities

Grok 4 is typically selected for:

  • Complex reasoning chains where intermediate constraints must be preserved.
  • Tool-calling assistants that execute multi-step workflows.
  • Coding and debugging use cases requiring strong analytical output.
  • Long-context tasks where large prompts or retrieval packs are required.
  • Real-time informed assistants when paired with xAI/X ecosystem integrations.

Its primary value is balancing frontier capability with strong tool use in an API-first setup, especially for products that benefit from xAI’s live-information positioning.

Technical Details

xAI’s API documentation and model catalog list Grok 4 with:

  • 256K context window.
  • High-tier pricing profile compared with Grok fast variants.
  • API-first access model with support for modern assistant/tool patterns.

xAI publicly documents context capacity but does not always separate a distinct “max generation” cap in the same way all providers do. In this profile, maxOutput tracks the documented upper-bound context ceiling for quick-reference consistency. In production, this means teams should enforce explicit output caps at the application layer rather than assuming provider-level defaults are consistent across all integration paths.

Pricing & Access

Listed API pricing (per 1M tokens):

  • Input: $3.00
  • Output: $15.00

Access options:

  • xAI API platform
  • Grok consumer surfaces and partner integrations depending on plan/tier

Because output pricing is significantly higher than input, use strict output-format constraints and routing guards in production.

Best Use Cases

Use Grok 4 for:

  • Advanced agent workflows with tool calls and error-recovery loops.
  • High-value coding/reasoning queries that justify premium rates.
  • Assistants that benefit from xAI ecosystem integration.
  • Analysis tasks where long context and directness are priorities.

Avoid using Grok 4 for simple, high-volume requests that cheaper variants can handle. If your product has large traffic spikes, combine Grok 4 with fallback tiers and strict timeouts so premium routes do not become a cost or latency bottleneck.

Comparisons

  • GPT-5.4 (OpenAI): Similar top-tier positioning; GPT-5.4 offers broader ecosystem maturity, while Grok can be attractive for xAI-native product stacks.
  • Claude Opus 4.6 (Anthropic): Opus is strong on conservative instruction adherence and enterprise governance; Grok emphasizes fast-moving frontier behavior and tool-centric usage.
  • Gemini 3.1 Pro Preview (Google): Gemini is stronger in built-in multimodal breadth; Grok is competitive for language-first reasoning and xAI platform integration.