Model Routing
Canopy auto-selects the best model for each task.
Это содержимое пока не доступно на вашем языке.
Canopy includes an intelligent model routing system that selects the most cost-effective model based on the task type.
How it works
When you start a task, Canopy evaluates its nature and recommends the appropriate model:
| Task Type | Recommended Model | Cost Profile |
|---|---|---|
| Scripts, one-liners, regex, SQL | DeepSeek V4 Flash / MiMo V2.5 | 🟢 Cheapest |
| General dev, refactoring, debugging | Kimi K2.6 | 🟡 Balanced |
| Investigation, large context, research | Qwen 3.7 Plus | 🟠 Capable |
| Architecture, planning, specs | GLM-5.1 | 🟣 Strategic |
| Complex coding | Kimi K2.7 Code | 🔴 Specialized |
Configuration
Set your default models in opencode.json:
{ "model": "kimi/k2.6", "small_model": "deepseek/v4-flash"}Benefits
- Cost savings — Simple tasks don’t pay for expensive reasoning models
- Speed — Scripts run on fast, lightweight models
- Quality — Complex tasks get the reasoning power they need
- Automatic — No manual switching required
Model routing in system prompts
Canopy’s system prompts include instructions for the AI to self-select the appropriate model. When the AI detects a task that would be better served by a different model, it can suggest switching.