Cloud provider services - verified October 18, 2026

Google Vector Search (Vertex AI / Gemini Enterprise Agent Platform)

Google Vector Search is listed under cloud provider services (deployment type: managed provisioned). First listed price, in its own unit: $3.00, per GiB processed for index build and update. Stated: hybrid and sparse search. Read from the vendor's own pages on October 18, 2026.

Read from docs.cloud.google.com, cloud.google.com. Docs - Pricing - Compare every vector database

Status

Availability label: Not stated
No GA or preview label is quoted from the pages checked.

Moved
Vertex AI Vector Search now sits under Gemini Enterprise Agent Platform
Google's Vector Search docs and pricing now live on Gemini Enterprise Agent Platform pages, and the pricing page adds an Agent Retrieval section priced per CU-hour and per read operation. The pages state no date for the change.
“Leveraging the ScaNN algorithm, Vector Search lets you build next-generation search and recommendation systems as well as generative AI applications.”
docs.cloud.google.com
“Storage-Optimized CU $2.30 / 1 hour”
cloud.google.com
Research note: Vertex AI documentation and pricing URLs now redirect to "Gemini Enterprise Agent Platform" pages (under docs.cloud.google.com/gemini-enterprise-agent-platform). Source page

How Google Vector Search charges

Every price Google Vector Search states, in its own unit, with the quote and where it comes from. Nothing is converted.

6 price rows in each product's own unit, grouped by unit and sorted only inside a unit
Price	Unit, as stated	Quote and source	Where it comes from
Other units, one row eachnot sorted: no two rows share a unit
$3.00	per GiB processed for index build and update	“Vector Search charges $3.00 per gibibyte (GiB) of data processed in all regions.” cloud.google.com	-
$0.45	per GiB ingested (streaming updates)	“Vector Search charges $0.45/GiB ingested for Streaming Update inserts.” cloud.google.com	-
Charged per node hour by machine type and region	per node-hour (index serving)	“The price corresponds to the machine type, by region, and is charged per node hour.” cloud.google.com	-
Per CU-hour
$2.30	per CU-hour (Agent Retrieval index serving, Storage-Optimized)	“Storage-Optimized CU $2.30 / 1 hour” cloud.google.com	Agent Retrieval sectionFrom the pricing page's new Agent Retrieval section, which it describes as a fully managed retrieval engine that unifies vector indexing and payload storage into a single service.
$0.065	per CU-hour (Agent Retrieval index serving, Performance-Optimized)	“Performance-Optimized CU $0.065 / 1 hour” cloud.google.com	Agent Retrieval sectionFrom the pricing page's new Agent Retrieval section, which it describes as a fully managed retrieval engine that unifies vector indexing and payload storage into a single service.
Per 100,000 read operations
$0.06	per 100,000 read operations (Agent Retrieval payload database)	“Read Operations $0.06 / 100,000 count” cloud.google.com	Agent Retrieval sectionFrom the pricing page's new Agent Retrieval section, which it describes as a fully managed retrieval engine that unifies vector indexing and payload storage into a single service.

Free tier and limits

Free tier: Not stated
No free tier is stated in the Vector Search section of the pricing page.
Limits: Shard sizes: 2, 20 or 50 GiB per shard
“SHARD_SIZE_SMALL 2 GiB per shard SHARD_SIZE_MEDIUM 20 GiB per shard SHARD_SIZE_LARGE 50 GiB per shard”
docs.cloud.google.com

Search and indexing

Max dimensions: Not stated
Not stated on the overview, index configuration or hybrid search pages checked.
Distance metrics: SQUARED_L2_DISTANCE, L1_DISTANCE, DOT_PRODUCT_DISTANCE, COSINE_DISTANCE
“SQUARED_L2_DISTANCE Euclidean (L2) Distance L1_DISTANCE Manhattan (L1) Distance DOT_PRODUCT_DISTANCE Default value.”
docs.cloud.google.com
“COSINE_DISTANCE Cosine Distance.”
docs.cloud.google.com
Index types: Tree-AH (ScaNN), Brute force
“algorithmConfig oneOf: - TreeAhConfig - BruteForceConfig”
docs.cloud.google.com
“Leveraging the ScaNN algorithm, Vector Search lets you build next-generation search and recommendation systems as well as generative AI applications.”
docs.cloud.google.com
Hybrid and sparse search: Yes: dense plus sparse embeddings in one index
“With hybrid search in Vector Search, you can mix both dense and sparse embeddings into a single vector index and run queries with dense embeddings, sparse embeddings, or both.”
docs.cloud.google.com
Metadata filtering: Not stated
Not described on the pages checked.
Multi-tenancy: Not stated
Not described on the pages checked.

Built-in embedding and reranking

Built-in embedding: Not stated
The Vector Search pages checked expect you to bring embeddings; Agent Retrieval pricing bills the embedding model separately.
Built-in reranking: Not stated
Not stated on the pages checked.

Deployment, license and SDKs

Regions and clouds: Google Cloud regions (region list on the pricing page)
“The following tables summarize the pricing of an index serving in each region where the Vector Search is available.”
cloud.google.com
BYOC and self-host: Not stated
Managed Google Cloud service; not applicable.
License: Not stated
Managed service; no license stated.
SDKs: Python, Java (client libraries)
“Agent Platform SDK for Python - Class MatchingEngineIndex - Class MatchingEngineIndexEndpoint Python client”
docs.cloud.google.com
“Java client - Class IndexServiceClient”
docs.cloud.google.com

Vendor-stated claims

Performance, scale and SLA figures as Google Vector Search states them. Not measured.

A minimal setup under $100 per month can handle high throughput for moderate-sized use casesvendor-stated

“Even a minimal setup (under $100 per month) can accommodate high throughput for moderate-sized use cases.”

docs.cloud.google.com

Research notes

Vertex AI documentation and pricing URLs now redirect to "Gemini Enterprise Agent Platform" pages (under docs.cloud.google.com/gemini-enterprise-agent-platform). The pricing page also lists "Agent Retrieval", described as "a fully managed retrieval engine that unifies vector indexing and payload storage into a single service", with CU-hour and per-operation pricing. The first node-hour table on the pricing page lists e2-standard-2 at $0.093808 per node hour, but the region that table applies to is not certain from the page text, so it is not recorded as a price.

FAQ

How much does Google Vector Search cost?

In Google Vector Search's own units: $3.00 (per GiB processed for index build and update); $0.45 (per GiB ingested (streaming updates)); Charged per node hour by machine type and region (per node-hour (index serving)); $2.30 (per CU-hour (Agent Retrieval index serving, Storage-Optimized)); $0.065 (per CU-hour (Agent Retrieval index serving, Performance-Optimized)); $0.06 (per 100,000 read operations (Agent Retrieval payload database)). As of October 18, 2026; prices are not converted.

What is the maximum vector dimension in Google Vector Search?

Not stated. Not stated on the overview, index configuration or hybrid search pages checked.

Does Google Vector Search support hybrid search?

Yes: dense plus sparse embeddings in one index, per the vendor: "With hybrid search in Vector Search, you can mix both dense and sparse embeddings into a single vector index and run queries with dense embeddings, sparse embeddings, or both."

Does Google Vector Search have built-in embedding or reranking?

Embedding: not stated. The Vector Search pages checked expect you to bring embeddings; Agent Retrieval pricing bills the embedding model separately. Reranking: not stated. Not stated on the pages checked.

Sources

Last verified 2026-10-18 from each vendor's own docs, pricing pages, READMEs and license files; every quote was string-matched against the page that day. This is a dated snapshot, not a benchmark: prices are shown in each product's own unit and change without notice, so check the linked source before you commit.

Newsletter

One email per week. New model releases, retrieval tooling and pricing changes and what they mean for what you build.