# MongoDB Atlas Vector Search — AI Compare Ninja

> Combine operational document data and vector search in the same MongoDB cluster.

## Overview
- **Category:** Development & Code (Databases, vector search & SQL)
- **Status:** Active · checked - Active Atlas Free, Flex, and Dedicated offers with integrated Vector Search.
- **Paid from:** $0.011 / $30 / Hourly (Atlas Flex)
- **Free access:** Free plan (512 MB, 100 operations/second, and 3 search/vector indexes.)
- **Website:** https://www.mongodb.com/products/platform/atlas-vector-search
- **Canonical URL:** https://www.aicompare.ninja/en/tool/mongodb-atlas-vector-search

## What it does
- Vector search, ANN, ENN, and hybrid search natively in the MongoDB cluster
- Native reranking for more precise search results
- Automated embeddings via Voyage models; one-time 200 million token allowance per organization
- Free M0 cluster: 512 MB, up to 100 operations/second, 3 vector indexes
- Flex cluster: 5 GB storage and up to 10 vector indexes for development
- Voyage-4 embeddings from 0.06 USD per million tokens; Voyage-4-lite from 0.02 USD

## In brief
MongoDB Atlas Vector Search integrates vector search directly into MongoDB. Operational data, metadata, and embeddings reside in the same cluster; ANN, ENN, hybrid search, native reranking, and automated Voyage embeddings are supported.

M0 Free provides 512 MB, 100 ops/s, and 3 vector indexes. Flex provides 5 GB for at most 30 USD per month. Automated Embedding includes 200 million tokens one-time; afterwards, Voyage-4 costs 0.06 USD, Voyage-4-lite 0.02 USD, and Voyage-4-large 0.12 USD per million tokens.


## Limits
Vector Search itself works on the Free cluster; its meaningful limits are 512 MB, 100 operations/second, and three search/vector indexes. Automated Embedding is billed separately by token; each organisation receives one-time 200M free tokens per model that do not renew. MongoDB says text sent to the service is used only to generate embeddings, is not stored, and is not used for model training. Embeddings are stored in the customer's own database. MongoDB hosts the models in a US Google Cloud region; cluster, egress, and optional Search Node costs remain separate.

## Pricing tiers
| Tier | Price | Period | What you get |
| --- | --- | --- | --- |
| Atlas Free | free | Permanent | 512 MB storage; Shared RAM and vCPU; Up to 100 operations/second; Maximum 3 search/vector indexes; 1 free cluster per project |
| Atlas Flex | $0.011 / $30 | Hourly | 5 GB storage; Shared RAM and vCPU with burst capacity; Up to 10 search/vector indexes; For development and testing; Usage based |
| Atlas Dedicated | $0.08 / $56.94 | Hourly | 10 GB to 4 TB storage; 2 to 768 GB RAM; 2 to 96 vCPUs; Production workloads; Optional dedicated search nodes |
| Automated Embedding | 0,02-0.18 $ per Million Tokens | Pay-as-you-go | One-time 200M free tokens per model and organisation; Voyage 4 Lite $0.02/M; Voyage 4 $0.06/M; Voyage 4 Large $0.12/M; Voyage Code 3 $0.18/M |

## Sources
All details regarding plans, limits, and features are based on official vendor documentation and the primary sources verified below.

- https://www.mongodb.com/pricing
- https://www.mongodb.com/docs/atlas/atlas-vector-search/tutorials/vector-search-quick-start/
- https://www.mongodb.com/docs/search/deployment/feature-compatibility/
- https://www.mongodb.com/docs/vector-search/crud-embeddings/automated-embedding/billing/
- https://www.mongodb.com/docs/vector-search/crud-embeddings/automated-embedding/overview/

---
*Source: AI Compare Ninja (2026) - https://www.aicompare.ninja/en/tool/mongodb-atlas-vector-search*