LLM cost dashboard: what to put on it
A good LLM cost dashboard answers two questions at a glance: where is the money going, and is anything about to break the budget. Most dashboards fail because they show one big number - total spend - and stop there. Total spend tells you that costs went up; it never tells you why.
What belongs on the dashboard
- Spend by dimension - broken out by team, product surface, environment, and model, not just one grand total.
- Cost per request and per successful task - the unit economics that reveal whether growth is efficient or just expensive.
- Cache-read ratio - the share of input tokens served from cache, the clearest signal of input-cost efficiency.
- Budget burn - spend against a monthly budget, with projected month-end based on run rate.
- Anomalies - sudden spikes by team or model, surfaced rather than buried in a chart.
Separate the estimate from the truth
An LLM cost dashboard has two layers, and conflating them is the most common mistake. Gateway and proxy logs give a near-real-time estimate - fast, useful for catching spikes the same day, but approximate. The provider invoice is the truth - slow, but authoritative. Show both, label which is which, and reconcile the estimate against the invoice each month so the dashboard keeps finance's trust.
Two audiences, two views
Engineers need request-level granularity to act - which endpoint, which model, which call pattern. Finance needs roll-ups they can put in a forecast - spend by team, trend, budget variance. A single dashboard can serve both, but design the views separately rather than forcing one audience to read the other's numbers.
Common mistakes that make a dashboard misleading
- Averaging cache-read and input tokens together, which hides the biggest cost lever.
- Tracking money as floating-point, which lets cent-level drift accumulate into reports.
- Showing spend with no owner - a dashboard with low attribution coverage looks precise but cannot drive a single decision.
- Reporting only lagging invoice data, so spikes are discovered weeks after they happen.
Related
- LLM cost monitoring - the monitoring practice behind the dashboard.
- LLM cost attribution - how the per-dimension breakdowns are built.
- Anomaly detection for AI spend - surfacing spikes automatically.
- LLM budget governance - turning the burn chart into budgets and alerts.
Want this applied to your own LLM spend? FinOps LLM runs a free audit of your AI costs and shows where the savings are. Book free audit →