Usage & cost

Token consumption and cost trends — segment by engine, model, agent, project, tool or skill to find what to optimize.

Cost · window——
Tokens · window——
Sessions——
Cache hit ratio—input that came from cache
Cost / 1k tokens—blended across models in window
Burn rate · 24h vs 30d—recent vs baseline

Cost & tokens over time

By engine

EngineSessionsTokensCostCache %

By model

ModelSessionsTokensCost$/sessp95 dur

By agent type

AgentSessionsTurnsCostTokens/turn

By project

ProjectSessionsTokensCost

By tool

ToolKindCallsSessionsCost share

By skill

SkillCallsTasksCost share

Top tasks by cost

TaskSessionsTurnsTokensCostDuration

Stop reason mix