Chapter 4: Model Providers and Runtime Strategy

Welcome to Chapter 4: Model Providers and Runtime Strategy. In this part of Strands Agents Tutorial: Model-Driven Agent Systems with Native MCP Support, you will build an intuitive mental model first, then move into concrete implementation details and practical production tradeoffs.

This chapter explains provider selection and runtime tuning decisions.

Learning Goals

• choose model providers based on constraints
• configure parameters for quality/cost/latency tradeoffs
• use provider abstractions cleanly
• avoid lock-in through adapter-friendly architecture

Provider Strategy

• start with one provider for baseline reliability
• use explicit model IDs and params in code
• benchmark task classes before multi-provider expansion

Source References

• Strands Model Provider Concepts
• Strands README: Multiple Model Providers
• Strands Custom Provider Docs

Summary

You can now make provider decisions that align with product and operations goals.

Next: Chapter 5: Hooks, State, and Reliability Controls

Source Code Walkthrough

Use the following upstream sources to verify model provider and runtime strategy details while reading this chapter:

• src/strands/models/ — the model provider implementations directory; each file implements the Model protocol for a specific provider (Bedrock, LiteLLM, Ollama, OpenAI-compatible).
• src/strands/models/bedrock.py — the Amazon Bedrock model provider, which is the default and most feature-complete provider implementation in the Strands SDK.

Suggested trace strategy:

• compare the __init__ signatures across model providers in src/strands/models/ to understand which parameters are provider-specific vs. universal
• trace how a model provider handles the stream method to understand the interface contract all providers must satisfy
• review src/strands/models/litellm.py to see how LiteLLM is used as a multi-provider gateway for non-Bedrock deployments

How These Components Connect

flowchart LR
    A[Agent configured with model provider] --> B[Model protocol interface]
    B --> C[Provider-specific implementation in models/]
    C --> D[API call to Bedrock / LiteLLM / Ollama]
    D --> E[Token stream returned to agent loop]