I looked into Genie Code costs in Databricks — it’s cheap until it isn’t

protmaks
Databricks MVP
I’ve been testing Genie Code in Databricks and wanted to understand not just the UX, but the actual cost behavior.
My impression so far:
  • for simple code edits / code assistance, it looks almost free,
  • but if Genie Code starts doing things that involve real compute, including cluster usage, the cost goes up quickly,
  • I saw scenarios that translate to about $3/hour,
  • and it’s not always obvious in billing where exactly that spend is coming from.

protmaks
Databricks MVP

Sorry, it's my bad, not 3$ but 1.5-2$

MrChromatic
New Contributor II

I was looking at the cost inferred by genie, but i'm not understanding how genie usage is calculated, i looked the system.billing.usage table, can you tell me if i can use this to filter genie usage.

It's almost free, the cost appears if it runs a cluster to check the values, but they cannot be linked to Genie Code

DoTA
New Contributor III

Great writeup - I'm also facing the same problem with Genie Code.

1. This billing attribution gap isn't specific to Genie Code, it shows up anywhere an agent/orchestration layer spins up ad-hoc compute (serverless one) on your behalf. Eventhough you can control this via budget policy. 

2. It is quite frustrating to not able to control the underlying model of Genie Code, I'm not even sure if Databricks is doing semantic model routing to cheaper model when the task is easy or not.

3. I'm also sensing that for complex task, Genie Code tend to perform this weird loop that caused so much money.

4. I wish they have a usage view such as /usage, /context of Claude code, PI, or KIRO.

 

The only 2 ways I'm able to control the cost of Genie Code is

1. Via serverless budget policy.

2. and the account level budget for Genie.