Retired May 2026
Grok 4
xAI's flagship reasoning model, released July 2025 and trained with reinforcement learning at pretraining scale on the Colossus cluster, aimed at complex multi-step tasks.
Key Specifications
- Context window
- 256K tokens
- Knowledge cutoff
- Released
- Inputs
- Text, image
- Capabilities Show details
- Reasoning Thinks before it answers, always on or as a switchable mode. Function calling Connect to external tools, APIs, and systems. Structured output Return responses in structured formats like JSON.
Frequently Asked Questions
What is Grok 4 good for?
Multi-step reasoning over text and images. It was xAI's flagship reasoning model from July 2025 until its retirement.
When is Grok 4 not a good fit?
Any new work: xAI retired it in May 2026, and its migration guide points to Grok 4.3 with low reasoning effort. No built-in web search or code execution, and closed weights.
How does Grok 4 compare with Grok 4.3?
xAI describes Grok 4.3 as a fast, reliable model with strong tool calling and instruction following, and it adds a configurable reasoning effort with a no-reasoning setting, which Grok 4 did not offer. Same text and image input. Grok 4 came out 9 months earlier and has a smaller context (256K against 1M tokens).
Can I self-host Grok 4?
No. xAI does not publish the weights; Grok 4 is available only through hosted APIs.
More from xAI
| Model | Context | Input /1M | Output /1M |
|---|---|---|---|
|
Grok 4.5
|
500K | $2.00 | $6.00 |
|
Grok 4.6
|
500K | $2.00 | $6.00 |
|
Grok 4.1 Fast (Non-Reasoning)
|
2M | ||
|
Grok 4.1 Fast (Reasoning)
|
2M | ||
|
Grok 3
|
131K | ||
|
Grok 3 Mini
|
131K | ||
|
Grok 4 Fast (Non-Reasoning)
|
2M | ||
|
Grok 4 Fast (Reasoning)
|
2M | ||