Model Routing
glossary beginner 3 min
Sources verified Dec 27, 2025
Choosing the right AI model for each task based on complexity and cost.
Simple Definition
Model Routing is deciding which AI model to use for a given task. Just like you wouldn't use a chainsaw to cut butter, you shouldn't use the most expensive AI model for simple tasks.
Technical Definition
The practice of directing prompts to appropriate models based on:
- Task complexity: Simple boilerplate vs. architectural decisions
- Cost constraints: High-volume tasks need cheaper models
- Latency requirements: Real-time features need faster models
- Quality requirements: Critical code may need more capable models
| Task Type | Model Class | Examples |
|---|---|---|
| Simple completion | Fast/cheap | GPT-3.5, Claude Haiku, local models |
| Complex reasoning | Capable/expensive | GPT-4, Claude Sonnet, Gemini Pro |
| Specialized tasks | Domain-specific | Code-specific models, fine-tuned models |
Key Takeaways
- Different tasks need different models
- Match model capability to task complexity
- Consider cost, latency, and quality tradeoffs
- Simple tasks often don't need the most expensive model
Sources
Tempered AI — Forged Through Practice, Not Hype
? Keyboard shortcuts