How context depth affects query cost
This guide explains how context depth affects query credit consumption, including how actions and model choice amplify the cost at higher tiers.
The more information your agent reads per question, the more query credits it uses. The cost scales with the tier you choose, and certain actions or premium models can amplify it further.
Base cost by tier
Each context depth option has a different query credit cost:
| Option | Query credits |
|---|---|
| Compact | 1 query credit |
| Balanced | 1 query credit |
| Extended | 2 query credits |
| Rich | 3 query credits |
| Max | 4 query credits |
How actions amplify cost
Certain actions cost more at higher context depths.
Verify Responses - Runs multiple checks per answer. At higher context depths, each check processes more content, so cost per answer increases.
Document Analyst - Processes attached files alongside your knowledge base. At higher context depths, combining file content with the broader context increases cost further.
Test at your intended depth before committing - the combined cost can be significantly higher than the base multiplier alone.
How model choice amplifies cost
Premium models (such as Claude Opus) cost more at higher context depths - the combined cost is higher than either increase on its own. Start at a lower depth and raise it only if quality requires it.
Managing cost
- Start lower, go higher as needed. Compact and Balanced cover most agents well. Only move to Extended or above when you have a specific reason - see Choose the right context depth for your use case.
- Lowering takes effect immediately. The context and history window shrink on the next message.
- Context depth is set per agent. Run different agents at different depths to match cost to the use case.
Note: If your query credits deplete faster than expected after changing context depth, check whether you also have Verify Responses or Document Analyst enabled.