Compare / S3 Vectors
WarpSearch vs Amazon S3 Vectors
Vector storage and similarity search inside S3. Cheap storage, no keyword or hybrid search.
The scoreboard
Price
Cheaper on 1 of 2 workloads, level on 1
Monthly, list prices. Bars are drawn to the middle of a range.
Performance
Likely faster
Likely faster, warm
AWS says "around 100 ms or less" for frequent queries and "under one second" for infrequent ones, as round numbers with no percentiles. Our warm vector median is 6 to 16 ms.
How they bill
| WarpSearch | S3 Vectors | |
|---|---|---|
| Billing unit | Flat $2 per million requests | $2.50 per million queries plus a per-TB fee on the index each query scans |
| Storage | $0.08 per GB-month, lower at volume | $0.06 per GB-month |
| Writes | $1.00 per GB | $0.20 per GB (128 KB minimum per request) |
| Keyword and hybrid search | Yes | No. Vectors with metadata filters only |
| Filterable metadata | Any mapped field | 2 KB per vector |
What S3 Vectors cannot do, or does only in part: Vector similarity search only: no keyword or hybrid search and no aggregations. Filterable metadata is capped at 2 KB per vector, and AWS positions it for infrequent queries.
Price
Monthly cost for four common workloads, at list prices in US East.
1M text chunks with 768-dim embeddings, 7 GB of JSON. 1M queries a month spread through the day. 1 GB written a month.
| WarpSearch | S3 Vectors | |
|---|---|---|
| Storage | $0.56 (7 GB held x $0.08) | $0.42 |
| Writes | $1.00 (1 GB x $1.00) | $0.20 |
| Search | $2.00 (1M requests) | $9.06 ($2.50 API + $6.56 data processed) |
| Data returned | $0 (about 40 GB, inside the free 100 GB) | $0 |
| Minimum | $10 minimum applies (usage is $3.56) | None |
| Total a month | $10 | $9.68 |
Vector search only. Text search needs a second system.
5M products with 384-dim embeddings, 10 GB of JSON. 50M queries a month (about 19 a second on average). 20 GB written a month.
| WarpSearch | S3 Vectors | |
|---|---|---|
| Storage | $0.60 (7.5 GB held x $0.08) | $0.60 |
| Writes | $20 (20 GB x $1.00) | $4 |
| Search | $100 (50M requests) | $956 ($125 API + $831 data processed) |
| Data returned | $18 (250 GB returned, 100 GB free, before compression) | $0 |
| Minimum | Cleared | None |
| Total a month | $139 | About $961 |
Vector search only.
10 TB of raw logs kept searchable. 1 TB written a month. 10,000 dashboard aggregation queries a month.
| WarpSearch | S3 Vectors | |
|---|---|---|
| Total a month | $1,465 | Not a fit |
No keyword search or aggregations.
1 PB of raw logs kept searchable. 50 TB written a month. 100,000 heavy aggregation queries a month.
| WarpSearch | S3 Vectors | |
|---|---|---|
| Total a month | $39,060 | Not a fit |
No keyword search or aggregations.
Performance
What S3 Vectors publishes
| Frequent queries | "Around 100 ms or less" |
|---|---|
| Infrequent queries | "Under one second" |
| Recall | "Over 90% average recall for most datasets" |
| Write limit | 2,500 vectors/s per index |
AWS gives round numbers, not percentiles: about 100 ms for frequent queries. Our measured vector median is 6 to 16 ms at 0.95+ recall on most datasets, warm.
What we measured
| Test | Result |
|---|---|
| Keyword search, 2.8M docs (LoTTE) | 2.1 ms median, 3,700 queries/s |
| Keyword search, 10M passages (MS MARCO v2) | 2.4 ms median (5.4 ms with an exact total hit count) |
| Keyword search, small BEIR sets | 1 to 6 ms median |
| Vector search, 10M x 128-dim (BigANN SIFT) | 0.964 recall@10 at 6.2 ms median, 162 queries/s |
| Vector search, 1M x 1536-dim (arXiv) | 0.961 recall@10 at 15.7 ms median |
| Filtered vector search, 10M (YFCC) | 0.959 recall@10 at 10.4 ms median |
| Log filters, 22M events (http_logs) | 1 to 2 ms median |
| Log aggregations, 22M events | 190 to 430 ms median (sort, daily histogram, avg by status) |
| Ingest, short text | 63,000 to 97,000 docs/s |
| Ingest, logs (247M events) | 76,000 to 81,000 events/s |
| Ingest, 768 to 1536-dim vectors | 2,200 to 4,000 vectors/s |
| Time to searchable | Seconds, or immediate with commit=true |
| Cold start | Not measured yet. First query on a 2.3 GB index: 17 to 19 ms vs 2.7 ms warm (page cache only) |
Measured October 2026 through our Elasticsearch-compatible API, warm cache, client in the same region, single runs. Every test ran on deliberately minimal hardware, 2 to 4 CPUs (a 4 CPU container or an AWS i8g.large with 2 vCPUs), so these are the least favorable results we can produce. More hardware only improves them. Numbers exclude internet round trips.
Bottom line
Choose WarpSearch for
- Keyword, vector and hybrid search in one index. S3 Vectors has vectors only.
- Query-heavy workloads: S3 Vectors charges per TB scanned, so the catalog workload costs about 7x ours.
- Lower latency on warm indexes.
Choose S3 Vectors for
- Small, rarely queried vector sets: $9.68 for the RAG workload vs our $10 minimum.
- Cheaper storage ($0.06 vs $0.08 per GB) and writes ($0.20 vs $1.00 per GB).
- Native S3 and Bedrock integration.
S3 Vectors is a cheap place to keep vectors you query now and then. Once you need keyword search, filters beyond 2 KB, or steady query volume, WarpSearch costs less and does more.
Sources
Data-processed fees are tiered by vector count: $0.004 per TB for the first 100K vectors, $0.002 up to 10M, $0.0004 above. Prices checked October 20, 2026. List prices in US dollars, before discounts. Estimates are marked. Tell us if something is out of date: contact us.