How to Choose an Embedding Model: Evaluation & Tradeoffs

info@qdrant.tech (Andrey Vasnetsov) — Tue, 15 Jul 2025 00:00:00 +0000

How to Choose an Embedding Model: Evaluation & Tradeoffs

How do you choose the right embedding model to achieve the best search quality? Public benchmarks such as the Massive Text Embedding Benchmark (MTEB) can help you narrow down the options, but their datasets rarely represent your own data. Search quality is not the only requirement, either: an accurate model may be too slow or expensive to run.

Selecting the best embedding model is a multi-objective optimization problem and there is no one-size-fits-all solution, and there probably never will be. In this guide, we will try to provide some guidance on how to approach this problem in a practical way, and how to move from model selection to running it in production.

Filtering in Vector Search: When to Use What

info@qdrant.tech (Andrey Vasnetsov) — Tue, 10 Sep 2024 00:00:00 +0000

Filtering in Vector Search: When to Use What

Imagine you sell computer hardware. A shopper looking for laptops under $1000 wants similar laptops, but a plain vector search can still return laptops over $1000. A payload filter guarantees that every result meets the constraint:

POST /collections/online_store/points/query
{
 "query": [0.2, 0.1, 0.9, 0.7],
 "filter": {
 "must": [
 { "key": "category", "match": { "value": "laptop" } },
 { "key": "price", "range": { "lte": 1000 } }
 ]
 },
 "limit": 3,
 "with_payload": true
}

This guide covers the decisions around a filter like this one: which fields to index, how Qdrant runs the search, and when to change the defaults. For the syntax of every filter condition, see the filtering documentation.

Search Design on Qdrant - Vector Search Engine

How to Choose an Embedding Model: Evaluation & Tradeoffs

How to Choose an Embedding Model: Evaluation & Tradeoffs

Filtering in Vector Search: When to Use What

Filtering in Vector Search: When to Use What