AWS S3 Storage Classes Reference

All S3 storage class names with retrieval time, cost tier and minimum duration.

Reference for Amazon S3 storage classes from Standard to Glacier Deep Archive with retrieval latency, durability, availability zones, minimum storage duration and cost tier. It runs free in your browser on Gera Tools, with nothing uploaded.

Last updated Source: Gera Tools

Which S3 class is cheapest for storage?

Glacier Deep Archive has the lowest per-GB storage price, designed for data accessed once or twice a year with retrieval times of up to 12 hours (standard) or 48 hours (bulk). It carries a 180-day minimum storage duration, so early deletes are still billed for the full period.

Amazon S3 storage classes

S3 storage classes let you trade retrieval speed for storage price. Picking the right one cuts the bill dramatically for cold data — but each colder class adds retrieval fees and minimum-duration charges. This reference covers every general storage class with its durability, availability-zone footprint, retrieval latency, minimum duration and relative cost tier.

How it works

All S3 classes share the same eleven-nines (99.999999999%) durability for multi-AZ classes; they differ on availability, retrieval and price. The cost model has three levers:

total cost = storage $/GB-month + request/retrieval fees + min-duration charge
  • Hot (Standard, Express One Zone): no retrieval fee, highest storage price, used for active data.
  • Infrequent (Standard-IA, One Zone-IA): lower storage price + a per-GB retrieval fee + 30-day minimum.
  • Archive (Glacier Instant, Glacier Flexible, Deep Archive): lowest storage price, retrieval from milliseconds to 48 hours, 90–180-day minimum.
  • Auto (Intelligent-Tiering): moves objects between tiers automatically, no retrieval fees, small monitoring charge per object.

Storage class comparison table

ClassAZ footprintRetrieval latencyMin storage durationBest for
Standard≥3 AZsMillisecondsNoneActive, frequently read data
Intelligent-Tiering≥3 AZsMilliseconds (frequent tier)NoneUnknown or changing access patterns
Standard-IA≥3 AZsMilliseconds (+ retrieval fee)30 daysBackups accessed monthly
One Zone-IA1 AZMilliseconds (+ retrieval fee)30 daysRe-creatable data, secondary copies
Glacier Instant Retrieval≥3 AZsMilliseconds (+ retrieval fee)90 daysLong-lived data accessed quarterly
Glacier Flexible Retrieval≥3 AZs1–12 hours (tiered)90 daysDR archives, long-term backups
Glacier Deep Archive≥3 AZsUp to 48 hours180 daysRegulatory retention, rarely accessed
Express One Zone1 AZSingle-digit milliseconds1 hourHigh-performance, latency-sensitive apps

Choosing the right class

If you read data daily or more: Standard is the baseline. The slightly higher storage price is offset by no retrieval fees, and you avoid any risk of minimum-duration overage.

If your access pattern is unpredictable: Intelligent-Tiering watches each object for 30 consecutive days of no access and moves it to the infrequent tier automatically, with no retrieval fee. It charges a small per-object monitoring fee, which makes it cost-ineffective for tiny objects (a few KB), but it is ideal for large media libraries or datasets with seasonal spikes.

If you know data will be accessed once a month or less: Standard-IA gives a meaningful storage price reduction. The catch is the 30-day minimum: deleting or transitioning an object before 30 days still charges for 30 days. This class is cost-effective only for data that genuinely sits untouched for months.

For archive data you might need in an emergency: Glacier Instant Retrieval provides archive-level storage pricing but millisecond access — closer to Standard-IA in retrieval behaviour but with a 90-day minimum. Ideal for medical images, media masters, or compliance records that are almost never opened but occasionally needed urgently.

For long-term backups measured in years: Glacier Flexible Retrieval or Deep Archive. Use Expedited retrieval on Flexible if you might need something within minutes; plan for 12-hour Standard retrieval for the normal case. Deep Archive’s bulk retrieval can take up to 48 hours — it is appropriate only for data where retrieval is a planned, rare event.

Tips and notes

  • Lifecycle rules transition objects between classes by age — never move tiny or short-lived objects to IA/archive; the minimum-duration charge eats the saving.
  • One Zone classes are not zone-redundant: use only for data you can regenerate.
  • Glacier Flexible Retrieval offers Expedited (1–5 min), Standard (3–5 h) and Bulk (5–12 h) retrieval options at different prices.
  • Deep Archive is the floor on storage price but the slowest to read — plan for up to 48-hour bulk retrieval.
  • Intelligent-Tiering has an optional archive activation for objects inactive for 90 or 180 days, extending savings further at the cost of longer retrieval.