Paste a line item from your invoice and find what it bills for, across the Google Cloud services covered so far.
Vertex AI Generative Model Tokens
Token charges for hosted generative models on Agent Platform (formerly Vertex AI), billed separately for input and output.
Google renamed Vertex AI to Gemini Enterprise Agent Platform in April 2026; earlier invoices and billing exports use the old name. Output tokens cost substantially more than input tokens, and agentic workloads invert the usual intuition by generating enormous input volume as they re-read their own context on every step.
Billed: Per 1,000,000 tokens, priced separately for input and output, with reasoning tokens billed as output. Output is typically several times the input rate, and some models charge more above a prompt-length threshold.
What triggers this charge
- Every prompt and completion, counted in tokens rather than characters or requests.
- Long context windows, where the entire prompt is re-billed on each call in a multi-step conversation.
- Agentic loops, which consume many times the tokens of a single chat exchange for the same task.
- Retries after a successful response: a request that returned 200 is billed even if your code discards the answer and asks again. Requests that fail with a 4xx or 5xx error are not charged.
- Reasoning ("thinking") tokens, which are billed at the output rate even though they never appear in the response.
How to reduce it
Turn on context caching
Cached input tokens are billed at a substantially lower rate. For agents and RAG systems that resend the same system prompt and documents, this is the largest single lever. Explicit caches also carry an hourly storage charge, so cache only what is reused.
Route by difficulty
Most requests do not need the largest model. Sending the easy majority to a smaller one and escalating only when needed changes the blended rate dramatically.
Use batch mode for anything not interactive
Batch and Flex requests are billed well below standard rates, and most enrichment and classification jobs do not need a synchronous answer.
Cap output length
Output tokens are the expensive half. Setting a sensible max output length bounds the worst case.
Rates
Sources and review
- Explanation reviewed:
- 1 Nov 2026, against Agent Platform generative AI pricing.
- Rates verified:
- not yet. No rate is shown on this page until one is loaded from the Cloud Billing Catalog API.
Now find out how much of this you are actually paying.
Finitizer breaks your Google Cloud bill down to the SKU, shows how each one has moved since last week, and separates rate changes from usage changes, so you know whether to talk to engineering or to procurement.
See how Finitizer works