Skip to content

03-model-and-provider-routing.md

Latest commit

 

History

History
60 lines (42 loc) · 1.76 KB

03-model-and-provider-routing.md

File metadata and controls

60 lines (42 loc) · 1.76 KB
layout default
title Chapter 3: Model and Provider Routing
nav_order 3
parent OpenCode Tutorial

Chapter 3: Model and Provider Routing

Welcome to Chapter 3: Model and Provider Routing. In this part of OpenCode Tutorial: Open-Source Terminal Coding Agent at Scale, you will build an intuitive mental model first, then move into concrete implementation details and practical production tradeoffs.

OpenCode is provider-agnostic by design. A strong routing strategy controls quality, cost, and latency.

Routing Strategy

Workload Recommended Model Class
repo analysis high-context reasoning model
code edits fast coding-focused model
long refactors stable high-accuracy model
follow-up fixes low-latency model

Practical Controls

  • use explicit model defaults for repetitive workflows
  • define fallback providers for outage resilience
  • separate experimentation profiles from production defaults

Common Failure Modes

  • oscillating output quality due to unpinned model selection
  • silent cost spikes from oversized model defaults
  • context-window mismatches for large monorepos

Source References

Summary

You now know how to build a provider strategy instead of relying on a single default model.

Next: Chapter 4: Tools, Permissions, and Execution

How These Components Connect

flowchart LR
    A[opencode config] --> B{Provider Router}
    B --> C[Anthropic Claude]
    B --> D[OpenAI / Azure]
    B --> E[Ollama Local]
    B --> F[Custom Provider]
    C --> G[Agent Response]
    D --> G
    E --> G
Loading