# Grok 4 (0709)
Specific dated release of Grok-4 from July 9, 2024. Premium baseline for complex synthesis, analysis, and instruction following.
## Model Details
| Property | Value |
|----------|-------|
| Model ID | `grok-4-0709` |
| Context Length | 256,000 tokens (256K input + 33K max output) |
| Knowledge Cutoff | November 2024 |
| Input Modalities | Text, Images |
| Output Modality | Text |
| Estimated Parameters | ~1.7 trillion |
## Pricing
| Type | Cost per 1M tokens |
|------|-------------------|
| Input | $3.00 |
| Output | $15.00 |
## Capabilities
- Text and image input support
- Parallel tool calling
- Structured outputs
- Real-time web search with X platform integration
- Voice mode with camera input
- Advanced semantic search on social data
## Supported Parameters
- `temperature`
- `top_p`
- `max_tokens`
- `tools`
- `tool_choice`
- `response_format`
- `seed`
- `logprobs`
- `top_logprobs`
## Tool Selection
Tool Selection Score: **0.85** - Consistently routes requests to correct APIs/databases.
## Usage
```python
from openai import OpenAI
client = OpenAI(
api_key=os.environ.get("XAI_API_KEY"),
base_url="https://api.x.ai/v1"
)
response = client.chat.completions.create(
model="grok-4-0709",
messages=[
{"role": "user", "content": "Analyze this complex problem"}
]
)
```
## Notes
- Reasoning is always enabled (no non-reasoning mode)
- Does NOT support `reasoning_effort` parameter
- Does NOT support `presencePenalty`, `frequencyPenalty`, or `stop` parameters
- Use real-time web search for information beyond knowledge cutoff
- Premium model for complex analytical tasks
