Overview
MongoDB Atlas is the managed version of MongoDB, run by MongoDB itself on AWS, Azure or Google Cloud in more than 125 regions. Every deployment is a three-node replica set, and dedicated clusters can span regions or even clouds. The same service adds Atlas Search, Vector Search, Stream Processing, Data Federation and an SQL interface.
There are three tiers. The Free cluster gives 512 MB on shared hardware with up to 100 operations per second, one per project, and no backups; it is meant for learning. Flex, which replaced the old M2 and M5 shared tiers and Serverless instances in 2025, bills hourly by throughput band and costs between about $8 and $30 a month.
Dedicated clusters start at M10 ($0.08 an hour, 2 GB RAM) and scale to hundreds of gigabytes of RAM and sharded clusters beyond 4 TB. MongoDB describes M10 and M20 as fit for pre-production and small workloads, and M30 and up, with dedicated vCPUs, for production. Storage, backups and data transfer add to the hourly price.
Atlas has no scale-to-zero and no database branching, and the hourly prices shown are defaults that change with provider, region and storage choice, so use the cost calculator in the Atlas console before committing.
Pricing and plans
Free tier
One Free (M0) cluster per project: 512 MB storage, shared RAM and vCPU, up to 100 operations per second, 500 connections; no backups, and idle clusters may be deactivated
At the limit Slows down Past 100 ops/sec or 10 GB of transfer a week the cluster is throttled; writes fail at 512 MB.
Paused if unused No backups| Plan | Price | What you get |
|---|---|---|
| Free (M0) | $0 | 512 MB storage, shared resources, 100 ops/second, one per project, no backups |
| Flex | $8 to $30/mo | $0.011/hour base for 0-100 ops/second, rising by tier to $30/mo at 400-500 ops/second |
| Dedicated M10 | $0.08/hr | 2 GB RAM, 2 vCPUs, 10-128 GB storage; M10 and M20 use shared vCPUs |
| Dedicated M30 | $0.54/hr | 8 GB RAM, 2 vCPUs, 40-512 GB storage; recommended for production traffic, supports sharding |
| Dedicated M50 | $2.00/hr | 32 GB RAM, 8 vCPUs, 160 GB-4 TB storage |
| Enterprise Advanced | Custom | Self-managed MongoDB Enterprise Server with Ops Manager and commercial license |
What it does
- Multi-cloud clusters
- Deploy on AWS, Azure or Google Cloud, including clusters that span regions or providers.
- Atlas Search
- Full-text search built on the same cluster, without syncing to a separate search engine.
- Vector Search
- Store embeddings next to operational data and run semantic search for retrieval-augmented generation.
- Flex tier
- Usage-banded pricing from $0.011 an hour for development and small apps, capped at $30 a month at 500 ops/second.
- Sharding and auto-scaling
- Dedicated tiers auto-scale cluster size and storage, and sharded clusters grow past 4 TB.
- Stream Processing and Data Federation
- Process event streams and query across Atlas and cloud object storage with the MongoDB query API.
Strengths and limitations
Strengths
- Permanent free cluster for learning and prototypes
- Choice of AWS, Azure or Google Cloud and many regions, including EU
- Search, vector search and streaming in the same managed service
- Clear path from Free to Flex to dedicated and sharded clusters
Limitations
- Free clusters have no backups and can be deactivated when idle
- Free and Flex clusters cannot write temporary files to disk for large aggregations
- No scale to zero on dedicated clusters, which bill every hour
- Prices vary by provider, region and storage, making estimates harder
- Document model; relational joins and SQL need the separate SQL interface
Who it suits
Good for
- Applications built on MongoDB's document model
- Teams that need search or vector search next to operational data
- Companies that want a specific cloud provider or multi-cloud setup
Look elsewhere if
- Relational workloads that rely on SQL joins and constraints
- Many small idle databases, where a scale-to-zero service costs less
Pricing history
-
M2/M5 and Serverless retired
Effective: Atlas stopped supporting M2/M5 and Serverless; M2/M5 clusters were migrated to Flex, and Serverless instances to Free, Flex or Dedicated depending on usage. Source ↗
-
Flex clusters replace M2/M5 and Serverless
From February 2025 (month only; no day given): new M2/M5 shared clusters and Serverless instances can no longer be created; Flex (from $0.011/hr, $8 to $30/mo) replaces them. Source ↗
Alternatives to MongoDB Atlas
| Tool | From | Free option | Stages |
|---|---|---|---|
| Crunchy Bridge Managed PostgreSQL on AWS, Azure and GCP | $9/mo | ||
| Render Managed PaaS | $7/mo | Free tier | |
| Aiven Managed open-source data platform (PostgreSQL, MySQL, Kafka, Valkey, OpenSearch, ClickHouse) | $5/mo | Free tier | |
| Redis Cloud Managed Redis | $5/mo | Free tier | |
| Supabase Managed Postgres backend platform | $25/mo | Free tier | |
| Convex Reactive backend and document database | $25/mo | Free tier |
Questions
Is MongoDB Atlas free?
Yes. One Free cluster per project with 512 MB of storage and up to 100 operations per second is free forever, but it has no backups.
What replaced MongoDB Atlas M2, M5 and Serverless?
Flex clusters. New M2, M5 and Serverless deployments stopped in February 2025, and remaining ones were migrated by January 22, 2026.
How much does a Flex cluster cost?
Flex starts at $0.011 an hour, about $8 a month, and rises with throughput to about $30 a month at 400 to 500 operations per second.
What is the cheapest dedicated Atlas cluster?
M10 at a base of $0.08 an hour with 2 GB RAM and 2 vCPUs; actual price depends on provider, region and storage.
Sources read for this page