Milvus vs. pgvector
Compare Milvus vs. pgvector for vector search workloads. We want you to choose the most suitable vector database for your use case, even if it’s not us.
Vector databases have become a core piece of infrastructure for modern AI applications, including retrieval-augmented generation (RAG), AI agents, multimodal and semantic search, and recommendation systems across a wide range of industries. Choosing the right vector database can directly affect the performance, scalability, cost, and reliability of these applications.
This Milvus vs pgvector comparison is written for engineers and technical teams evaluating vector databases for real-world, production workloads. While Milvus and pgvector both support basic vector search capability, they differ significantly in areas such as architecture, scalability, performance, reliability, and many other areas. These differences often remain subtle during early experiments but become increasingly important as data volume grows, workloads diversify, and systems move from prototype to production.
The goal of this guide is not just to list features, but to help you determine which system is better aligned with your specific use case, constraints, and growth stage—even if our product (Milvus / Zilliz Cloud) is not the final choice.
Milvus vs pgvector at a Glance
Vector database designed for AI-native workloads, supporting large-scale vector storage and similarity search
No. pgvector is just a vector search add-on to Postgres
Apache 2.0
PostgreSQL License (similar to MIT)
41,924
18,973
Milvus is an open-source, high-performance, and highly scalable vector database that can store, index, and search billion-scale unstructured data through high-dimensional vector embeddings. It is perfect for building modern AI applications such as retrieval augmented generation (RAG), semantic search, multimodal search, and recommendation systems. Milvus runs efficiently across various environments, from laptops to large-scale distributed systems. It is available as open-source software and a cloud service.
pgvector is an extension for PostgreSQL that adds support for vector similarity search directly within the database. It allows developers to store, index, and query vector embeddings alongside relational data. pgvector is ideal for hybrid applications that combine traditional relational queries with vector-based retrieval, leveraging PostgreSQL’s mature ecosystem.
Benchmarking Milvus and pgvector Using Your Datasets
Don’t take any vendor’s word for performance — test it yourself.
VectorDBBench is an open-source benchmarking tool built specifically for comparing vector databases under fair and reproducible conditions. It lets you measure real performance — latency, throughput, recall, indexing speed, and scaling behavior — using the same workloads across different systems or even using your own datasets.
This makes it easy to see how Milvus and pgvector actually behave in practice, not just in marketing materials. Every benchmark can be reproduced locally, in your own environment, so you can validate the results that matter for your application.
Check out the VectorDBBench Leaderboard for a quick look at mainstream vector DB performance.
Why 10,000+ Enterprises Teams Switch to Milvus and Zilliz Cloud
Most vector databases look fine in demos or small-scale deployment, but the gaps show up in production—when datasets grow, embeddings refresh frequently, and latency must stay stable under real traffic. That’s where teams turn to Milvus and Zilliz Cloud (managed Milvus service).
Milvus is an open-source, high-performance vector database used by 10,000+ enterprise teams worldwide and trusted by a large open-source community with 41K+ GitHub stars. It handles tens of millions to tens of billions of vectors, frequent inserts and deletes, and hybrid search (vector + keyword + metadata + reranking) without disruptive reindexing or fragile tuning. Performance stays predictable as data volume, query patterns, and model embeddings evolve. This is why Milvus is widely deployed for enterprise RAG, AI agents, semantic and multimodal search, and recommendation systems—workloads where instability becomes visible immediately.
Zilliz Cloud delivers the same Milvus architecture as a managed service, with an advanced vector engine (Cardinal) for higher performance, plus elastic scaling, high availability, enterprise-grade security and compliance, and global deployment. Teams get production-ready reliability without operating or babysitting the database.
- Compare Milvus and Zilliz Cloud with any other vector database
- Benchmark Milvus or Zilliz Cloud yourself using VectorDBBench
Milvus’s real advantage was how easy and friendly it made things to understand and execute."
Hanlian Lyu
a Product Owner and BI Expert at Volvo Cars
Milvus has done an extraordinary job in revolutionizing Likee's video deduplication system, which significantly fueled the growth of BIGO's short-video business."
Xinyang Guo
Software Engineer at BIGO
Milvus has dramatically facilitated the MMU team in building various business systems and effectively supports our rapid business growth. Thanks to the Milvus team for developing such a fantastic vector database with stable vector search capabilities and rich functionalities."
The MMU team
Shopee
Our search system has been much more intelligent, stable, and reliable using Milvus. "
Rahul Yadav
Software Engineer at Tokopedia
We plan to expand the use of Milvus in different fields like content moderation and restriction and customized video services. BIGO and Milvus working together will benefit both businesses and I look forward to Milvus and its community to keep growing and prosper."
Xinyang Guo
Software Engineer at BIGO
I was benchmarking different vector databases, and Milvus was at the top among all the popular options on the market."
Saumil Patel
the Senior Data Scientist from Ivy.ai
We've gained so much from the Milvus community that we decided to contribute features like "hot reload," which have also benefited our internal operations."
Dennis Zhao
AI Infrastructure Lead at SmartNews
We tested every mainstream vector database, and Milvus delivered the best overall performance."
Su Wei
CTO of Shining
Milvus delivers unparalleled performance and flexibility, integrating seamlessly with leading vector index libraries like Faiss. Its intuitive API and robust solutions for high availability make it an indispensable tool in our APK security efforts."
Wei Huang
Senior Research Engineer
Milvus consistently outperformed Weaviate, emphasizing the indexing time for scenario S9, closely resembling the FARFETCH product catalog's dimensions."
PEDRO MOREIRA COSTA
Applied Scientist
When you're engineering a solution as complex as ours, you're not just ticking boxes—you're looking for that sweet spot where all your must-haves intersect. Think of it as an eight-circle Venn diagram; while many databases met one or two of our criteria, Milvus was the only one sitting right at the intersection of all eight. It checked every single box for us—something no other solution managed to do."
Jack Fischer
Co-founder & CTO at Credal AI
Milvus has not only streamlined but also remarkably expedited the retrieval of millions of semantic vectors, showcasing a nearly tenfold improvement compared to our previous experience with other vector similarity search engines."
Tingting Wang
NLP Algorithm Engineer at Sohu
Milvus searches tens of millions of vectors in milliseconds, providing optimal performance while keeping development costs low and resource consumption minimal. "
Yu Fang
AI scientist at Mozat
Milvus-powered vector search has been running steadily in our recommendation systems, providing high performance and allowing us more flexibility in selecting models and algorithms."
VIPSHOP Search Service Team
VIPSHOP
Zilliz Cloud stood out for us with its comprehensive range of index types, automatically optimizing for the perfect balance between recall and performance. Its robust security and private networking features, combined with the advantage of a fully-managed database, have significantly lightened our operational load, enabling us to concentrate on driving innovation rather than managing backend complexities. "
Ben Kramer
Co-founder & CTO of Monterey AI
When we conducted our first test run (on Zilliz Cloud), the performance improvement was astonishing. Our search time went from eight seconds down to sub one second. We were nearly falling off our chairs with amazement at the speed."
Alex Alexander
Co-founder & CEO of Picdmo Inc.
You can think of LlamaIndex as a black box around your data and LLM."
Jerry Liu
Co-founder and CEO of LlamaIndex
Zilliz Cloud perfectly aligns with MindStudio's vision. Its high-performance, secure platform and multi-tenancy simplifies data management and unlocks unprecedented productivity and innovation for our users’ AI applications. "
Sean Thielen
CTO @ MindStudio
As our business has expanded, the demands on our vector database have increased. We need a solution that minimizes operational costs, offers elastic scaling capabilities to manage vast amounts of vector data and unexpected traffic surges, provides faster vector search speeds, and ensures a high service level agreement (SLA)."
Chenhui Li
Tech Lead at Shulex
I appreciated using the open standard evaluation benchmarks for machine learning in general; this is also true for vector databases. The ones that Zilliz often publicizes have been beneficial, and the fact that they are open is significant. "
Sam Butler
Director of Machine Learning @ Dopple.AI
I can set up Milvus for scale in an hour, not overthinking it, and have a clear architecture path forward to grow my use case much larger without having to redo everything."
Zen Yui
Co-founder & CTO at Troop
Reliability, scalability, and performance are great. The fact that we can outsource all this kind of administration management and don't have to worry about servers breaking helps us put resources into product development and innovation, which truly differentiates our business. "
Michal Oglodek
CTO of Ivy.ai
Since transitioning from the open-source Milvus vector database to the fully managed Zilliz Cloud, we’ve experienced significant improvements in business performance. We’ve achieved lower operational costs, increased search speed, a more flexible system architecture, and a more stable user experience. Zilliz Cloud also provides expert support to resolve issues quickly and effectively. Overall, Zilliz Cloud has given us greater convenience and a competitive edge, and we are very pleased and optimistic about this change."
Shengyi Pan
CTO of Shulex
Even with numerous concurrent searches, we didn’t notice any slowdown in search speed with Milvus."
Mr. Zhang
BOSCH’s principal software engineer
We need an indexing technology that can handle complex search requirements and generative models, reduce training costs, improve update efficiency, and adapt flexibly to evolving data and query needs."
Mr. Zhang
BOSCH’s principal software engineer
When we identify a need for specific data, we can often find the required data in our database the same day using text or image search with Milvus. This greatly improves our data processing efficiency and has a positive effect on our business operations."
Mr. Zhang
BOSCH’s principal software engineer
I really like how Milvus' hybrid search allowed me to blend semantic and keyword search, which is crucial in a domain as technical and complex as EU policy."
Alessandro Saccoia
Co-Founder @ Veridien.ai
Milvus is renowned as one of the most advanced vector database platforms for AI applications. Rakuten Symphony engineers identified the Milvus Vector Database - an open source database which is horizontally scalable - as their platform of choice for LLM use and developing and maintaining AI applications."
Rakuten Symphony Engineering Team
Zilliz Cloud gave us the speed and scale we needed to power visual search at Leboncoin, meeting our sub-200ms latency target and making product discovery seamless for millions of users."
Yann Lemonnier
ML Engineer
Zilliz is a very integral part of our workflow. If we just swapped Zilliz Cloud with something else, the kind of loops that we get might not make sense, which means the end composition might not sound very nice or might not be very accurate to the text prompt you had given."
Sangarshanan Veera
Senior Software Engineer at Beatoven.ai
Zilliz Cloud enabled us to scale search while dramatically reducing our costs. Its hybrid search capabilities helped us achieve higher relevance for both text and image queries."
Celine Lightfoot
CTO of Beni
The competitive advantage for meeting transcription products in the future will come from personalization – providing tailored insights and recommendations based on each user's specific meeting data and usage patterns. This is why performant vector search is imperative."
TJ Dai
CTO of Notta
After integrating Zilliz Cloud vector database service, our system performance has significantly improved. During implementation, the Zilliz Cloud expert team provided excellent support and assistance, giving our EviMed platform a strong competitive advantage in the industry."
Dr. Zeyuan Wang
CEO of EviMed
During batch ingestion tests, Milvus demonstrated that it could complete an entire collection dump into the database at speeds 5–10 times faster than competitors."
Team Lead
AI, ML, and Platform Solutions
By migrating to Zilliz Cloud, we've reduced our vector database infrastructure costs by approximately 70% compared to our self-hosted setup. This allows us to reinvest those savings into improving our core AI capabilities rather than managing database infrastructure."
Nguyễn Ngọc Hải Đăng_ Nguyễn Nhật Khoa
AI Engineer at CX Genie
Thanks to the well-designed Python SDK and REST API, we were able to integrate Zilliz Cloud with our LangChain-based architecture in a matter of days. The schema-based collections perfectly aligned with how we structure our data, making the transition nearly seamless."
Nguyễn Ngọc Hải Đăng_ Nguyễn Nhật Khoa
AI Engineer at CX Genie
With Zilliz Cloud, we've achieved query latencies as low as 5-10ms across our million-vector database. This represents performance that's twice as fast as our previous solution, which directly translates to more responsive chatbots for our customers."
Nguyễn Ngọc Hải Đăng_ Nguyễn Nhật Khoa
AI Engineer at CX Genie
We have achieved a true consciousness of data... bringing the data together in the way that an individual doing their job needs to see it."
Nathan Morris
Co-Founder of Filevine
Working with Zilliz Cloud has been transformative for our AI agent architecture. The hybrid search capability alone delivered a 40% accuracy improvement, and the scalability means we never worry about performance, even during the highest traffic periods. It's been essential to our continued growth."
Sasidhar Janaki
Senior Software Engineer at Rexera
When it comes to vector databases, Milvus has impressed us with its performance and scalability, meeting our stringent criteria for handling our AI use case backlog."
Team Lead
AI, ML, and Platform Solutions
We were facing latency issues and scaling challenges with our previous solutions. When traffic spiked with millions of customer requests, our self-hosted infrastructure couldn't keep up, and document retrieval was taking too long. "
Sasidhar Janaki
Senior Software Engineer at Rexera
As a client of a law firm, at some point clients will demand that their law firms use AI because it's going to make them a better attorney to have the curation done versus some of the practices now where they just don't have complete info."
Brianna Connelly
AI Team Lead at Filevine
We have the full end-to-end data... law as a timeline. We have the incident, the start, everything that's happening, and the conclusion of the case. There's an enormous amount of power in having that all within the work system."
Brianna Connelly
AI Team Lead at Filevine
Milvus scaled really well with batches ranging from 1,000 to millions of records. That really impressed me."
Todor Voynikov
Data Engineer at TrialHub
The sub-10 millisecond latency for queries is quite the benchmark in the industry for vector databases. Combined with the cost savings and reliability, Zilliz Cloud has become a strategic enabler for our broader vision of transforming conversational AI across industries."
Rachit Jindal
Senior AI Engineer at Verbaflo.ai
We were running a multilingual RAG system at scale, indexing all of Wikipedia into tens of millions of high-dimensional vectors. Latency targets were tight. We needed a system that could handle real-time retrieval over millions of knowledge vectors without breaking under load. Zilliz gave us that. It freed up engineering cycles and let us focus on improving reasoning on the model side, not managing infrastructure. That reliability mattered because we were operating at 3,500+ dimensions per vector, across a fast-growing corpus, with sub-300ms latency requirements in production. "
Dr. Pratyush Kumar
Co-Founder of Sarvam
When it comes to vector databases, Milvus has impressed us with its performance and scalability, meeting our stringent criteria for handling our AI use case backlog."
Toby Yu
Team Lead of AI, ML, and Platform Solutions
During batch ingestion tests, Milvus demonstrated that it could complete an entire collection dump into the database at speeds 5-10 times faster than competitors."
Toby Yu
Team Lead of AI, ML, and Platform Solutions
We've got millions of monthly active users and all of the underlying data when we're trying to go find related conversations, find updates to an action item, find referenced documents...Milvus serves as the central repository and powers our information retrieval among billions of records."
Rob Williams
Co-Founder and CTO at Read AI
What we wanted was to push intelligence to the user before they even asked. Milvus is what made that viable."
Rob Williams
Co-Founder and CTO at Read AI
Our strongest competitive moat isn't our AI models—it's our ability to deploy those models at scale with an exceptional user experience. Zilliz Cloud gave us that capability."
Ethan Zheng
Co-Founder & CTO of Jobright.ai
It has saved us cost by at least 25 to 30% on what we were earlier incurring on the vector database, with potential savings reaching up to 40-50% during peak traffic periods. "
Rachit Jindal
Senior AI Engineer at Verbaflo.ai
Choosing Zilliz Cloud was one of our best early decisions. It enabled us to build the product we envisioned rather than the product our infrastructure limitations would allow. In AI applications, that difference often determines success or failure."
Ethan Zheng
Co-Founder & CTO of Jobright.ai
From a system stability perspective, it's really quite good. Over the year-plus that we've been using it—from version 2.4.3 to now 2.5.8—I honestly haven't encountered many issues. The system can just run there for months, with new data being written every day and users searching every day, without any problems. I don't need to worry about it."
Team Lead
Senior Infrastructure Engineer
With Zilliz Cloud, we moved from operating at our limits to building with confidence. It gave us the scale, performance, and flexibility to protect music rights in real time—something we couldn’t achieve with traditional systems."
George Kastrinakis
Director of Data Science and AI Services at Orfium
Among Milvus's rich vector search capabilities, features such as support for multiple ANN index types, multi-vector support, and hybrid search have proven especially valuable in real-world service environments. As Milvus continues to evolve with new capabilities, NAVER expects even broader applications across its services."
NAVER Engineering Team
It was the best thing to offer—performance-wise, cost-wise, and ease-of-use-wise."
George Kastrinakis
Director of Data Science and AI Services at Orfium
Milvus transformed our ability to detect semantic plagiarism at scale. We can now process variable workloads ranging from 10 to 10,000+ documents daily while maintaining cost-effectiveness, which would have been impossible with traditional solutions."
Teis Petersen
Engineering Team Lead, UNIwise
Whenever we demonstrate our solution with Milvus, we’re effectively crushing the cloud vendor solution’s performance. It’s a great benchmark because most enterprise users already know the vendor solution, so the comparison is immediate."
Hanlian Lyu
a Product Owner and BI Expert at Volvo Cars
If I were to choose again, I would still choose Milvus at this point. The scalability, documentation quality, and continuous innovation make it the right foundation for our plagiarism detection platform."
Teis Petersen
Engineering Team Lead, UNIwise
From a system stability perspective, it's really quite good. Over the year-plus that we've been using it—from version 2.4.3 to now 2.5.8—I honestly haven't encountered many issues. The system can just run there for months, with new data being written every day and users searching every day, without any problems. I don't need to worry about it."
Jianping Wang
NVIDIA engineer
Milvus has become the bridge that connects our multi-modal foundation models with real-world applications. It's not just about performance – it's about enabling entirely new approaches to biological discovery that were previously impossible."
Xiaoming Zhang
VP of Technology at Biomap
Milvus has become the only technical choice for vector databases in our upcoming business expansion across the life sciences industry."
Xiaoming Zhang
VP of Technology at Biomap
Milvus’s real advantage was how easy and friendly it made things to understand and execute."
Hanlian Lyu
a Product Owner and BI Expert at Volvo Cars
Milvus has done an extraordinary job in revolutionizing Likee's video deduplication system, which significantly fueled the growth of BIGO's short-video business."
Xinyang Guo
Software Engineer at BIGO
How to Migrate to Milvus/Zilliz
Migrating to Milvus or Zilliz Cloud is simple. You can import data from Qdrant, Weaviate, Pinecone, Elasticsearch, OpenSearch, Amazon S3 Vectors, PostgreSQL, and more using built-in tools that automate extraction and loading.
For production workloads, we support zero-downtime migration with live data sync. Many teams cut their vector infrastructure cost by up to 50% after switching, while gaining faster performance and more predictable scaling.
Start Migrating to Milvus/Zilliz now
Ready to migrate your unstructured and vector data? Whether you're moving from Elasticsearch, Pinecone, or another database, Zilliz makes it easy.
Don't just take our word for it
Zilliz is named a Leader in the Forrester Wave™ Vector Database Report
Learn more
SOC 2 Type II
Security and organizational controls for cloud providers.
ISO/ICE 27001
Global standard for information security management systems.
GDPR
Privacy protections for EU and EEA data.
HIPAA
U.S privacy regulation safeguarding health information.
Check our Trust Center to learn how Zilliz meets the highest standards of security and compliance.
Compare Milvus with others
The Definitive Guide to Choosing a Vector Database
Overwhelmed by all the options? Learn key features to look for & how to evaluate with your own data. Choose with confidence.