Skip to content

Latest commit

 

History

History
43 lines (36 loc) · 1013 Bytes

model-selection.md

File metadata and controls

43 lines (36 loc) · 1013 Bytes

Selecting the Right Model

Selection Criteria

  1. Task Type:

    • Chat/conversation
    • Text completion
    • Code generation
    • Embeddings

  2. Model Characteristics:

    • Context window size
    • Training data recency
    • Special capabilities (function calling, etc.)

  3. Performance Needs:

    • Response time
    • Output quality
    • Cost considerations

  4. Rate Limits:

    • Requests per minute/day
    • Token limits
    • Concurrent requests

Recommendation Process

  1. Define your use case
  2. Identify must-have features
  3. Compare models in playground
  4. Test with your specific prompts
  5. Evaluate performance

Common Models

  • GPT-4 variants
  • Claude models
  • Open-source alternatives
  • Specialized models (code, etc.)

Model Card Checklist

  • Review the model card in the Marketplace for:
    • Supported capabilities (chat, code, vision, embeddings)
    • Rate limits and pricing
    • Example code and SDK support
    • Provider-specific features (function calling, streaming, etc.)