Costs

Where the LLM budget goes.

Which feature is burning the budget?

Projected /mo
$541
from last 24h × 30
Top feature
chat_summary
46% of spend · $249 /mo
Top spender
u_heavy_19
3.6% of spend
Heavy 5%
53%
47 users drive this share
By featureBy userBy modelWhich feature is burning the budget?
FeatureShare of spendCallsUsersAvg in / outTop model24hProjected /mo
chat_summary
46%
84394621 / 394claude-opus-4-7$8.31$249.21
code_assist
24%
57271754 / 666claude-opus-4-7$4.35$130.37
support_bot
15%
11948332 / 244claude-opus-4-7$2.77$83.02
data_extraction
10%
71311390 / 1420claude-sonnet-4-6$1.81$54.25
search_qa
4%
9040958 / 273gpt-4o$0.800$24.01
moderation
0%
6028168 / 65gpt-4-mini$0.004$0.115

Knowing where the money goes is half the answer.

Insights → turns these breakdowns into concrete actions: model swaps, prompt-caching opportunities, output caps, and fine-tune ROI estimates.