Mini tools| Model | Input | Cached input | Output | Unit |
|---|---|---|---|---|
|
Grok 4.2
xAI
|
$1.25 | $0.2 | $2.5 | per 1M tokens |
|
Grok 4.3
xAI
|
$1.25 | $0.2 | $2.5 | per 1M tokens |
|
Grok 4.5
This model
SpaceX
|
$2 | $0.5 | $6 | per 1M tokens |
|
Grok Build 0.1
xAI
|
$1 | $0.2 | $2 | per 1M tokens |
Grok 4.5
By SpaceX
Model family: Grok
Grok 4.5 is a mixture-of-experts model trained on trillions of tokens capturing developer and agent interactions with codebases and tools, then refined with reinforcement learning on difficult problems across software engineering and broader knowledge work. Unlike its coding-specialist predecessor, it was trained on a deliberately broader data mix including STEM tasks and research papers, giving it proficiency across domains beyond coding. It is built to handle long-running tasks requiring tool use, problem investigation, error recovery, and result verification. Base pricing is $2 per million input tokens and $6 per million output tokens; a faster variant is priced at $4/$18 per million tokens.
New
Text
Overview
Grok 4.5 is a mixture-of-experts language model built to handle difficult, long-running tasks that require creatively using tools, across software engineering, data science, finance, and legal work. It uses reasoning and tool use to investigate problems, recover from mistakes, and verify results.
Pricing
Compare Grok 4.5 with other models listed in the same vendor pricing tiers and context lengths.
Tier
Standard
Priority
| Model | Input | Cached input | Output | Unit |
|---|---|---|---|---|
|
Grok 4.2
xAI
|
$2.5 | $0.4 | $5 | per 1M tokens |
|
Grok 4.3
xAI
|
$2.5 | $0.4 | $5 | per 1M tokens |
|
Grok 4.5
This model
SpaceX
|
$4 | $1 | $12 | per 1M tokens |
|
Grok Build 0.1
xAI
|
$2 | $0.4 | $4 | per 1M tokens |
About SpaceX
Industry:
Aerospace Technology
Company Size:
13000
Location:
Hawthorne, California, US
Website:
spacex.com
Tools using Grok 4.5
-
Grok — v4.5Handles difficult long-running tasks more effectively by investigating problems, using tools, recovering from mistakes, and verifying results in realistic environments. Extends beyond software engineering into data science, finance, legal work, and other computer-based knowledge tasks with a broader training mix. Improves agent-style computer work through training on large volumes of codebase interactions and developer-agent behavior with software tools. Gains stronger robustness on hard problems through reinforcement learning environments designed specifically to remain challenging even for frontier systems. Offers a separate fast variant, indicating a higher-speed option alongside the main weight class for lower-latency workflows.
Last updated: July 8, 2026
