The /groq command enables you to interact with Groq’s high-speed language models. Perfect for:

• Fast text generation
• Document analysis
• Complex reasoning tasks
• Multi-file processing
• JSON-structured outputs

​

Basic Usage

Use the command to interact with Groq models:

/groq analyze this document and summarize key points

/groq use model llama-3.3-70b to write a technical report

/groq process these files and extract data as JSON

​

Key Features

​

Model Selection

• Multiple model options
• Varying context lengths
• Different speed tiers
• Specialized capabilities
• Automatic model selection

​

File Processing

• Direct file URL handling
• Multi-file support
• No content retrieval needed
• Efficient processing
• Batch operations

​

Output Formats

• JSON (default)
• Plain text
• Markdown
• HTML
• Custom formats

​

Example Commands

​

Basic Query

/groq explain quantum computing in simple terms

​

With Model Selection

/groq use deepseek-r1-distill-llama-70b to solve this complex math problem

​

Process Files

/groq analyze these documents [file1.pdf, file2.txt] and compare their content

​

Custom Output Format

/groq format as markdown: create a table comparing these products

​

Long Context Task

/groq use llama-4-maverick-17b with 128k context to analyze this book

​

Available Models

​

Fast & Efficient

• gemma2-9b-it: Very fast, 8K context
• llama-3.1-8b-instant: Very fast, 131K context
• llama-4-scout-17b: Very fast, 16K context

​

Balanced Performance

• llama-3.3-70b-versatile: Fast, 32K context
• mixtral-8x7b-32768: Fast, 32K context
• llama-4-maverick-17b: Fast, 131K context (default)

​

Advanced Reasoning

• deepseek-r1-distill-llama-70b: Complex problem solving, 8K context
• llama-3.1-70b-versatile: Long context versatile tasks, 131K context

​

Model Selection Guide

Choose based on your needs:

• Quick responses: Use instant or scout models
• Long documents: Use maverick or llama-3.1 models
• Complex reasoning: Use deepseek model
• General purpose: Use versatile models
• Default choice: llama-4-maverick-17b

​

Output Control

​

JSON Format (Default)

/groq extract product details from this description

​

Plain Text

/groq format as plaintext: write a story about robots

​

Markdown

/groq format as markdown: create documentation for this API

​

Custom System Prompt

/groq with prompt "You are a data scientist" analyze this dataset

​

Tips

• Default model handles most tasks well
• Use specific models for specialized needs
• JSON output is default for structured data
• Include file URLs directly in commands
• Specify format explicitly when needed