Skip to main content
Watch: Stop Building Apps. Build Tools for AI Agents

Open source, self-hosted - verified

Milvus (open source)

Milvus is listed under open source, self-hosted (deployment type: open source self-host). Open source; no price. Managed Milvus is offered as Zilliz Cloud (separate entry). Max dimensions: 32,768. Stated: hybrid and sparse search and built-in embedding. Read from the vendor's own pages on October 18, 2026.

Read from milvus.io, raw.githubusercontent.com. Docs - License - Compare every vector database

Status

Availability label
Not stated

No release-status label is quoted from the docs pages checked; the docs version selector lists v3.0.x, v2.6.x, v2.5.x and v2.4.x.

How Milvus charges

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

Open source; no price. Managed Milvus is offered as Zilliz Cloud (separate entry).

Free tier and limits

Free tier
Not stated

Open-source software; no hosted free tier applies.

Limits

Up to 1,024 partitions per collection

“Number of resources in a collection Resource Limit --- --- Partition 1,024”
milvus.io

Search and indexing

Max dimensions

32,768

“Dimensions of a vector Property Limit --- --- Dimension 32,768”
milvus.io
Distance metrics

L2, IP, COSINE, Jaccard, Hamming

“- Euclidean distance (L2) - Inner product (IP) - Cosine similarity (COSINE)”
milvus.io
“- Jaccard - Hamming - BIN_FLAT - BIN_IVF_FLAT”
milvus.io
Index types

FLAT, IVF_FLAT, IVF_SQ8, IVF_PQ, IVF_RABITQ, HNSW, HNSW_SQ, HNSW_PQ, HNSW_PRQ, DISKANN, SCANN, AISAQ, FAISS, GPU_CAGRA, GPU_IVF_FLAT, GPU_IVF_PQ, GPU_BRUTE_FORCE

“FLOAT_VECTOR - FLAT - IVF_FLAT - IVF_SQ8 - IVF_PQ - IVF_RABITQ - HNSW - HNSW_SQ - HNSW_PQ - HNSW_PRQ - DISKANN - SCANN - AISAQ - FAISS - GPU_CAGRA - GPU_IVF_FLAT - GPU_IVF_PQ - GPU_BRUTE_FORCE”
milvus.io
Hybrid and sparse search

Yes: BM25 full-text search combinable with dense vector search

“Milvus provides full text search powered by the BM25 relevance algorithm”
milvus.io
“By integrating full text search with semantic-based dense vector search, you can enhance the accuracy and relevance of search results.”
milvus.io
Metadata filtering
Not stated

Scalar filtering is not described on the specific pages checked (limits, metrics, index overview, full-text search, multi-tenancy).

Multi-tenancy

Database, collection, partition or partition-key level

“Milvus supports multi-tenancy at four levels: Database, Collection, Partition, and Partition Key.”
milvus.io

Built-in embedding and reranking

Built-in embedding

Yes, via the Function module calling external embedding providers (Milvus 2.6.x)

“The Function module in Milvus allows you to transform raw text data into vector embeddings by automatically calling external embedding service providers”
milvus.io
Built-in reranking
Not stated

Not stated on the pages checked (limits, metrics, index overview, full-text search, multi-tenancy, embedding function and SDK install pages).

Deployment, license and SDKs

Regions and clouds
Not stated

Self-hosted software; regions do not apply.

BYOC and self-host

Open source (Apache-2.0)

“Apache License Version 2.0, January 2004”
raw.githubusercontent.com
License

Apache-2.0

“Apache License Version 2.0, January 2004”
raw.githubusercontent.com
SDKs

Python (PyMilvus), Java, Go, Node.js

“Install SDKs - PyMilvus - Java SDK - Go SDK - Node.js SDK”
milvus.io

Vendor-stated claims

Performance, scale and SLA figures as Milvus states them. Not measured.

No performance, scale or SLA figure is recorded from the pages checked.

Research notes

The Embedding Function page is marked "Compatible with Milvus 2.6.x" while the docs version selector also offers v3.0.x.

FAQ

How much does Milvus cost?

Open source; no price. Managed Milvus is offered as Zilliz Cloud (separate entry).

What is the maximum vector dimension in Milvus?

32,768, per the vendor: "Dimensions of a vector Property Limit --- --- Dimension 32,768"

Does Milvus support hybrid search?

Yes: BM25 full-text search combinable with dense vector search, per the vendor: "Milvus provides full text search powered by the BM25 relevance algorithm"

Does Milvus have built-in embedding or reranking?

Embedding: Yes, via the Function module calling external embedding providers (Milvus 2.6.x). Reranking: not stated. Not stated on the pages checked (limits, metrics, index overview, full-text search, multi-tenancy, embedding function and SDK install pages).

Sources

Last verified 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

Know when a vector database changes its pricing

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