Pinecone Pricing: Serverless Storage, Read and Write Units (1 October 2026)
Independent reference for what Pinecone charges to store and query the embeddings you generate. Every figure below comes from Pinecone's own pricing page or cost documentation, with the source linked.
How much does Pinecone cost?
On the Standard plan Pinecone charges $0.33/GB/month for storage, $16-$18 per million read units and $4-$4.50 per million write units, with a $50/month minimum. The Starter plan is free up to 2 GB of storage, and Builder is $20/month flat for up to 10 GB.
Source: https://www.pinecone.io/pricing/, read 1 October 2026.
Published Plan Pricing
| Starter | Builder | Standard | Enterprise | |
|---|---|---|---|---|
| Price | Free | $20/month flat | $50/month minimum | $500/month minimum |
| Storage | Up to 2 GB | Up to 10 GB | $0.33/GB/month | $0.33/GB/month |
| Read units | Up to 1M/month | Up to 2M/month | $16-$18 per million | $24-$27 per million |
| Write units | Up to 2M/month | Up to 5M/month | $4-$4.50 per million | $6-$6.75 per million |
| Egress | Up to 1GB/month | Up to 10GB/month | $0.10/GB, 100GB/month included | $0.10/GB, 100GB/month included |
| Indexes | Up to 5 | 10 per project | 20 per project | 200 per project |
| Namespaces per index | 100 | 1,000 | 100,000 | 1,000,000 |
Source: pinecone.io/pricing, read 1 October 2026. Standard includes a 3 week trial with $300 in credits. The Standard and Enterprise minimums are usage minimums: Pinecone says that once usage exceeds the minimum, you pay as you go.
| Data operation (Standard, Enterprise) | Price |
|---|---|
| Import from object storage | $0.25 per GB |
| Backups | $0.10/GB/month |
| Restore from backup | $0.15 per GB |
How Read and Write Units Are Counted
Pinecone's cost documentation (read 1 October 2026) ties read cost to namespace size. A query uses 1 RU for every 1 GB of namespace size, with a minimum of 0.25 RUs per query. An upsert uses 1 WU for each 1 KB of the request, with a minimum of 5 WUs per request, and an upsert that overwrites an existing record also pays for the existing record's size.
The practical consequence: query cost grows with the size of the namespace you search, not only with query volume. Splitting a large corpus into per-tenant namespaces keeps each query's read-unit charge small. Shrinking vectors does the same, which is where dimension reduction pays twice on Pinecone: once on storage and again on every query.
Worked Example: 1M Vectors on Standard
This is our arithmetic on Pinecone's published rates, using raw vector size only, so metadata stored alongside each record adds to it. At this scale storage and reads come to about $12 a month, so the $50 Standard minimum is what you actually pay. A corpus this size also fits Builder's 10 GB allowance at $20/month, provided monthly reads stay within its 2M read units.
Pinecone vs Qdrant vs Weaviate vs Self-Hosting
The three managed vendors bill on different meters, so a per-GB comparison only tells part of the story.
| Billing meter | Free tier | Paid entry point | |
|---|---|---|---|
| Pinecone | Storage per GB + read and write units | Starter: up to 2 GB | Builder $20/month flat; Standard $50/month minimum |
| Qdrant Cloud | vCPU, RAM and disk, billed hourly | 0.5 vCPU / 1GB RAM / 4GB disk | Standard tier, usage-based |
| Weaviate Cloud | Vector dimensions + storage per GiB | 100,000 objects, 1 GB memory | Flex from $45/month |
| Self-hosted | Your own servers | Open-source Qdrant, Weaviate, pgvector | Server cost plus operations time |
Pinecone is the only one of the three that charges per query, which suits spiky or low traffic and penalises heavy query load against a large namespace. Qdrant and Weaviate charge for the cluster or the stored dimensions whether you query or not. Past a few tens of millions of vectors with steady traffic, the self-hosting break-even is worth running. The storage comparison puts all five options side by side at 1M, 10M and 100M vectors.
Pod-Based Indexes
Older Pinecone pricing guides quote fixed monthly pod prices. Pinecone's pod documentation (read 1 October 2026) now describes pod-based indexes as legacy: customers who signed up for a Standard or Enterprise plan on or after August 18, 2025 cannot create them. Pod prices no longer appear on the pricing page, so new projects should be costed on the serverless rates above.
Frequently Asked Questions
How much does Pinecone cost?
What does a Pinecone read unit cost?
How many read units does a Pinecone query use?
Is there a free Pinecone tier?
Can I still use Pinecone pod-based indexes?
Sources
- Plan prices, storage, read and write units, egress, import, backup, restore, index limits: https://www.pinecone.io/pricing/, read 1 October 2026.
- Read and write unit consumption per request: https://docs.pinecone.io/guides/manage-cost/understanding-cost, read 1 October 2026.
- Pod-based index status: https://docs.pinecone.io/guides/indexes/pods/understanding-pod-based-indexes, read 1 October 2026.