AI Models on Vectly

Vectly provides access to multiple state-of-the-art AI models, each with unique strengths and optimal use cases. This guide helps you choose the right model for your needs.

Available Models

Claude Models (Anthropic)

Claude 4 Opus

• Best for: Complex reasoning, detailed analysis, creative writing
• Strengths:
  • Exceptional reasoning capabilities
  • Nuanced understanding of context
  • Strong safety features
  • Great for technical documentation
• Context Window: 200K tokens
• Credit Cost: Higher (see pricing)

Claude 4 Sonnet

• Best for: Balanced performance and cost
• Strengths:
  • Good reasoning abilities
  • Faster than Opus
  • Cost-effective for most tasks
  • Reliable for code generation
• Context Window: 200K tokens
• Credit Cost: Medium

GPT Models (OpenAI)

GPT-4.1 Turbo

• Best for: General-purpose tasks, code generation, analysis
• Strengths:
  • Excellent code understanding
  • Strong general knowledge
  • Fast response times
  • Good at following instructions
• Context Window: 128K tokens
• Credit Cost: Medium-High

GPT-4o

• Best for: Multimodal tasks, vision understanding
• Strengths:
  • Can analyze images
  • Native multimodal understanding
  • Efficient processing
  • Good for visual content
• Context Window: 128K tokens
• Credit Cost: Medium

GPT-4o Mini

• Best for: Quick queries, simple tasks, testing
• Strengths:
  • Very fast responses
  • Low credit usage
  • Good for simple questions
  • Ideal for development/testing
• Context Window: 128K tokens
• Credit Cost: Low

O1 Models (OpenAI)

O1 Preview

• Best for: Complex reasoning, mathematical problems, coding challenges
• Strengths:
  • Advanced reasoning capabilities
  • Step-by-step problem solving
  • Excellent for STEM tasks
  • Self-correcting reasoning
• Context Window: 128K tokens
• Credit Cost: Very High

O1 Mini

• Best for: Reasoning tasks with budget constraints
• Strengths:
  • Good reasoning abilities
  • More affordable than O1 Preview
  • Faster processing
  • Suitable for moderate complexity
• Context Window: 128K tokens
• Credit Cost: High

Choosing the Right Model

By Use Case

📝 Writing & Content Creation

• Best: Claude 4 Opus
• Alternative: GPT-4.1 Turbo
• Budget: Claude 4 Sonnet

💻 Code Generation & Debugging

• Best: GPT-4.1 Turbo
• Alternative: Claude 4 Sonnet
• Budget: GPT-4o Mini

🧮 Complex Problem Solving

• Best: O1 Preview
• Alternative: Claude 4 Opus
• Budget: O1 Mini

🔍 Research & Analysis

• Best: Claude 4 Opus
• Alternative: GPT-4.1 Turbo
• Budget: Claude 4 Sonnet

🖼️ Image Analysis

• Best: GPT-4o
• Alternative: GPT-4.1 Turbo (text only)

⚡ Quick Questions

• Best: GPT-4o Mini
• Alternative: Claude 4 Sonnet

Model Features

Reasoning Mode

Available for Claude and O1 models:

• Visible Reasoning (Claude): See the AI's thought process
• Hidden Reasoning (O1): Advanced reasoning without visible steps

Web Search

Compatible with all models:

• Search current information
• Verify facts and data
• Find recent events
• Additional credit cost applies

Prompt Caching

Reduces costs for repetitive contexts:

• Available for Claude models
• Automatically caches repeated prompts
• Significant savings for long contexts

Credit Usage

Credit consumption varies by:

1. Model Selected: Premium models cost more
2. Input Length: Longer prompts use more credits
3. Output Length: Longer responses cost more
4. Features Used: Web search, reasoning add costs

Approximate Credit Costs (per 1K tokens)

Best Practices

1. Start Small

• Test with GPT-4o Mini first
• Upgrade to premium models if needed
• Use prompt caching for repeated contexts

2. Match Model to Task

• Don't use O1 Preview for simple questions
• Don't use Mini models for complex analysis
• Consider context window limits

3. Optimize for Credits

• Be concise in your prompts
• Use system prompts effectively
• Enable caching when applicable
• Monitor usage in real-time

4. Leverage Strengths

• Use Claude for nuanced writing
• Use GPT-4 for code and general tasks
• Use O1 for complex reasoning
• Use Mini models for iteration

Model Comparison Table

Coming Soon

We're constantly adding new models:

• Gemini Pro models
• Specialized fine-tuned models
• Custom model deployment
• Model mixing capabilities

Stay tuned for updates!