model-selection

What problem does it solve?

This Skill guides the complex process of selecting and managing LLM models and providers, addressing challenges like cost overruns, performance bottlenecks, and vendor lock-in. It ensures optimal model choice for specific tasks, improves reliability through fallbacks, and helps control operational expenses.

Core Features & Use Cases

• Model Registry: Centralizes metadata for various LLMs (Anthropic, OpenAI, Google), including capabilities, pricing, and recommended use cases.
• Model Router: Implements logic to dynamically route prompts to the most appropriate model based on task complexity, length, or specific requirements.
• Fallback Chain: Provides patterns for implementing robust fallback mechanisms, ensuring continuous service even if a primary model or provider fails.
• Cost Optimization: Offers tools to estimate and analyze LLM costs, helping identify the cheapest model for a given task and compare expenses across different models.
• Use Case: An LLM application needs to handle diverse user requests, from simple Q&A to complex code generation. This skill helps implement a ModelRouter to send simple queries to a fast, cheap model (e.g., Haiku) and complex requests to a flagship model (e.g., Sonnet), while a FallbackChain ensures resilience.

Quick Start

Set up a ModelRegistry with Claude Sonnet and Haiku models, including their capabilities and pricing.