Vertex AI / Gemini cost monitoring

See what your Gemini usage costs, per model

CostMon reads Vertex AI spend from your GCP billing export. It joins that spend with Cloud Monitoring's per-model token counts. A model that costs more each day appears on the chart quickly, not in next month's GCP invoice.

Start free See how it works

GCP billing export + Cloud Monitoring token usage, live today

Vertex AI spend comes through the same GCP billing export as every other Google Cloud service. It is easy to lose inside one "GCP" total. Even when someone splits it out by service, the export does not show which model drives it. Batch prediction and context caching both lower the price. They are opt-in. Nothing in the billing console says when you miss them.

Cost anatomy

What actually drives your Vertex AI bill

Per-token, priced by model tier: Vertex AI / Gemini bills per input and output token, separately, and the rate varies widely by model tier. The cheapest and most capable Gemini models can differ several-fold on the same million tokens.
Long-context surcharge: Requests above certain context-length thresholds (e.g. 200K input tokens on some models) bill at a surcharge over the standard token rate. A workload that grows its prompts over time can cross that line without anyone noticing.
Batch vs. online serving: Batch and Flex prediction cost about half as much as standard online serving for the same model. A nightly summary job that runs online leaves that discount unused. CostMon cannot see which mode a call used.
Context caching, its own line item: Context caching bills cached input tokens far below the standard input rate. It adds a separate cache-storage charge. Without it, a long repeated prompt pays full input price on every call. CostMon reads input and output token counts only.
Same billing export as the rest of GCP: Vertex spend arrives in the same BigQuery billing export as all other Google Cloud services. Unless you split it out by service, it blends into the "GCP" total.

Alerts

How CostMon warns you about Vertex AI spend

CostMon's default rule watches every service on your account from the day you connect it. An alert always shows up in your CostMon feed. Add an email, Slack, or webhook channel to send it somewhere else too.

Anomaly rulesource = vertex, service = gemini-1.5-pro, metric = tokens

A workload's token use climbing day over day, for example as its prompts grow.

Budgetsource = vertex

A monthly budget on total Vertex AI spend catches it quietly growing inside the wider GCP bill.

Sample alert, illustrative

From: alerts@costmon.com
Subject: CostMon alert: Anomaly: vertex / gemini-1.5-pro spend spike

Anomaly: vertex / gemini-1.5-pro spend spike

vertex / gemini-1.5-pro was $230.00 on 2026-07-19 — 77% above its 14-day average of $130.00.

Severity: warning
Occurred on: 2026-07-19

To find what to cut, the recommendations list flags idle spend and commitment gaps to review.

Why CostMon

Built for Vertex AI spend, from day one

Per-model attribution, broken out from the GCP line

CostMon joins your GCP billing export with Cloud Monitoring's per-model token metrics. Vertex AI spend breaks down by model and by day. It no longer hides in the rest of your GCP bill.

Spot the model that costs more each day

The daily view shows spend and token counts for each model. A model with a steep rise stands out. Then you check that workload for a cheaper way to run it, such as Batch.

Joins Vertex AI spend to Bedrock and Anthropic

Gemini spend usually hides inside the GCP invoice. CostMon splits it out and puts it beside your Bedrock and Anthropic spend. You can compare AI vendors on one chart.

Read-only, billing-export and Monitoring-scoped credentials

CostMon connects with a service account scoped to bigquery.dataViewer, jobUser, and monitoring.viewer only. It's read-only and never touches your Vertex AI models, prompts, or infrastructure.

Per-tokeninput and output priced separately, and separately by model tier
~50%typical Batch API discount vs. on-demand for the same model
Billing export + MonitoringGCP spend joined to per-model tokens, daily

Getting started

From API key to first alert, for Vertex AI

01
Connect a read-only GCP service account
Grant a service account bigquery.dataViewer and jobUser on your billing export dataset, plus monitoring.viewer for token metrics. CostMon never requests access to your Vertex AI models or infrastructure.
02
CostMon joins billing export and token usage
CostMon filters Vertex AI rows out of your billing export. It matches each day's rows to Cloud Monitoring token metrics, split into input and output per model. CostMon then saves the joined rows as standard cost entries.
03
See the surcharge, then get the alert
Follow Gemini spend for each model, one day at a time. As prompts grow, the daily token total climbs and a rule fires.

FAQ

Common questions about the Vertex AI connector

How does CostMon connect to Vertex AI?

It uses the same read-only service account as the GCP connector. That account has bigquery.dataViewer and jobUser on your billing export dataset. It also has monitoring.viewer for Cloud Monitoring's token metrics. CostMon never requests access to your Vertex AI models or prompts.

Is Vertex AI spend separate from the rest of my GCP bill?

No. It comes from the same GCP billing export as everything else. CostMon filters it to Vertex AI and generative AI services. It joins the rows with per-model token metrics. You need no separate connector or invoice.

Can CostMon tell me if I should be using the Batch API?

Not directly. CostMon shows daily spend and token counts for each model. A high-volume model stands out there. The token metric covers online serving only, so CostMon cannot tell if a call used Batch. It also does not move the workload for you.

Does CostMon track context caching?

No. CostMon reads two token directions from Cloud Monitoring: input and output. It has no count of cached tokens. It does show the spend for each model, so a cost rise on a model is visible.

What does the Vertex AI connector cost?

A flat monthly rate per plan. It never charges a percentage of your Vertex AI spend or the savings CostMon helps you find. Free includes 2 connectors; Professional and Enterprise include unlimited connectors.

Connectors

Cloud, AI, and SaaS: all in one place

AWS, GCP, and Azure for cloud. Anthropic, OpenAI, Amazon Bedrock, Vertex AI / Gemini, and Helicone for AI and LLM spend. Snowflake and Databricks for the data cloud. Datadog, GitHub, and Vercel for the rest of the stack. Add CSV import for any tool without a native connector. That's 14 connectors, all normalized into the same unified view.

Connect it. Get the first alert.

Connect a provider and CostMon builds a daily baseline for it. From there, a spike reaches your team as an alert.

Start free See how to connect a provider