The /anthropic command enables you to interact with Claude AI models for various tasks. Perfect for:

• Complex reasoning and analysis
• Code generation and review
• Image understanding
• Research and writing
• Creative tasks

​

Basic Usage

Use the command to interact with Claude:

/anthropic ask Claude "Explain quantum computing in simple terms"

/anthropic analyze this image [upload image] and describe what you see

/anthropic using claude-opus-4 help me solve this complex problem: ...

​

Key Features

​

Text Generation

• Ask questions
• Generate content
• Analyze text
• Summarize documents
• Translate languages

​

Image Analysis

• Describe images
• Extract text from images
• Analyze visual content
• Answer questions about images

​

Model Selection

• Choose specific models
• Optimize for speed or quality
• Balance cost and performance

​

System Instructions

• Set custom behaviors
• Define output formats
• Specify response styles

​

Example Commands

​

Basic Question

/anthropic what are the main differences between Python and JavaScript?

​

With System Prompt

/anthropic system "You are a helpful coding assistant" review this code for bugs: ...

​

Image Analysis

/anthropic analyze [image] and identify all objects in the scene

​

JSON Response

/anthropic system "Respond only in JSON format" list 5 European capitals

​

Complex Task

/anthropic using claude-opus-4 analyze this research paper and provide key insights

​

Available Models

​

Premium Models

• claude-opus-4-20250514: Complex reasoning, research
• claude-sonnet-4-20250514: Coding, analysis, production

​

Standard Models

• claude-3-5-sonnet-20241022: General chat, summaries, code (default)
• claude-3-5-haiku-20241022: Fast chat, simple queries

​

Legacy Models

• claude-3-opus-20240229: Complex reasoning
• claude-3-sonnet-20240229: Balanced performance
• claude-3-haiku-20240307: Ultra-fast, basic tasks

​

Tips

• Default model is claude-3-5-sonnet for balanced performance
• Use Opus models for complex reasoning tasks
• Use Haiku models for simple, fast responses
• Images should be in JPEG, PNG, GIF, or WebP format