Loading...
Live

Track spend, burn rate, and budgets across all compute providers (Modal, OpenAI, Anthropic, GCP)

Providers
Historical Spend
Cost by Source

Daily Spend (Last 30 Days)

Spend Projection (Next 60 Days)

Time range:

Breakdown by Provider

Modal, OpenAI, Anthropic, GCP — billing API + estimates. How is this calculated?

ProviderCostRecords% of TotalStatusSource
GCP Breakdown

By Project

ProjectCostRecords

By Service

ServiceCostRecords
Proxy Inference

Cost by Model

ModelCostCallsAvg Cost / Call

Recent Calls

TimestampModelCostTokens
Extended Views

Spend by Team

TeamCostRecords% of Total

Spend by Benchmark

BenchmarkCostRecords% of Total

Provider Detail

How costs are calculated

Records with source_type=billing_api come from provider billing APIs. Records with source_type=estimate are computed from token counts and pricing tables. Records with source_type=proxy_spend_log are per-call costs captured by the LiteLLM proxy.