Compare / OpenSearch Serverless
WarpSearch vs Amazon OpenSearch Serverless
AWS's serverless OpenSearch. You pay for compute by the hour (OCUs) while it is awake. WarpSearch bills each request instead.
The scoreboard
Price
Cheaper on 2 of 4 workloads
OpenSearch Serverless is cheaper for: Mid-size logs, Large logs.
Monthly, list prices. Bars are drawn to the middle of a range.
Performance
No like-for-like data
No published latency to compare
AWS publishes no query latency, vector recall or indexing rate for Serverless. It does publish a 10 to 30 second wake from zero on NextGen. We have not measured our own cold start yet, so we make no speed claim here.
How they bill
| WarpSearch | OpenSearch Serverless | |
|---|---|---|
| Billing unit | Per GB stored, per GB written, per request | Per OCU-hour (6 GB RAM slice) for indexing and search, plus storage |
| Smallest awake cost | $10 a month minimum | $350 a month (2 search OCUs at $0.24/hour) while traffic keeps it awake |
| Scale to zero | After 15 minutes idle | NextGen only, after 10 minutes idle (fixed). Classic: 2 OCU minimum |
| Logs (time series) | Yes | Classic only. NextGen does not support time-series collections |
| API | Elasticsearch, OpenSearch and Solr compatible | OpenSearch API, some APIs unsupported |
| Hybrid search | Keyword and vector in one index | Yes |
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 | OpenSearch Serverless | |
|---|---|---|
| Storage | $0.56 (7 GB held x $0.08) | $3 (10 GB x $0.30 hot storage) |
| Writes | $1.00 (1 GB x $1.00) | $4 if batched daily, $175 if streamed all day |
| Search | $2.00 (1M requests) | $350 (2 OCUs awake all month) |
| Data returned | $0 (about 40 GB, inside the free 100 GB) | About $0.50 |
| Minimum | $10 minimum applies (usage is $3.56) | None |
| Total a month | $10 | $357 to $530 |
1M queries a month is one every 2.6 seconds, so it never idles 10 minutes.
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 | OpenSearch Serverless | |
|---|---|---|
| Storage | $0.60 (7.5 GB held x $0.08) | $4.50 |
| Writes | $20 (20 GB x $1.00) | $175 to $350 (steady updates keep 1 to 2 OCUs awake) |
| Search | $100 (50M requests) | $350 to $700 (2 to 4 OCUs) |
| Data returned | $18 (250 GB returned, 100 GB free, before compression) | About $22 |
| Minimum | Cleared | None |
| Total a month | $139 | $550 to $1,080 |
10 TB of raw logs kept searchable. 1 TB written a month. 10,000 dashboard aggregation queries a month.
| WarpSearch | OpenSearch Serverless | |
|---|---|---|
| Storage | $455 (8.5 TB held: 1 TB x $0.08, 7.5 TB x $0.05) | $240 (10 TB x $0.024, Classic) |
| Writes | $1,000 (1 TB x $1.00) | $175 (1 OCU) |
| Search | $10 (10,000 requests, about 10 CPU-seconds each) | $175 to $350 (1 to 2 OCUs) |
| Data returned | $0 | Negligible |
| Minimum | Cleared | 2 OCUs, covered |
| Total a month | $1,465 | $590 to $765 |
Classic time-series collection. NextGen cannot hold logs.
1 PB of raw logs kept searchable. 50 TB written a month. 100,000 heavy aggregation queries a month.
| WarpSearch | OpenSearch Serverless | |
|---|---|---|
| Storage | $31,280 (850 TB held, tiered) | $24,000 (Classic) |
| Writes | $6,450 (1 TB x $1.00, 9 TB x $0.25, 40 TB x $0.08) | Not published; we estimate $2,500+ |
| Search | $1,330 (100,000 requests, about 133 CPU-seconds each) | $2,400+ (at least 14 OCUs) |
| Data returned | $0 | Negligible |
| Minimum | Cleared | Covered |
| Total a month | $39,060 | About $27,000 to $30,000 |
Estimate. AWS does not publish OCU throughput. Classic caps hot storage at 10 TiB per collection.
Performance
What OpenSearch Serverless publishes
| Query latency | Not published |
|---|---|
| Wake from zero (NextGen) | About 10 seconds; docs say 10 to 30 seconds. Requests are queued |
| Time to searchable | About 10 seconds, not configurable |
| Vector latency and recall | Not published |
| Indexing throughput per OCU | Not published |
AWS publishes no query latency for OpenSearch Serverless, so a number-for-number comparison is not possible. What is published: a 10 to 30 second wake from zero, and about 10 seconds before writes are searchable.
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
- Small and bursty apps: $10 instead of $350+ once traffic keeps OpenSearch awake.
- Catalog search: about 4 to 8x cheaper at 50M queries a month.
- Logs and app search on the same scale-to-zero pricing. NextGen cannot hold logs, and Classic keeps 2 OCUs running.
Choose OpenSearch Serverless for
- Mid-size logs: Classic is about half our price, because our writes cost $1.00 per GB for the first TB.
- Petabyte logs: Classic storage is $0.024 per GB against our $0.035 top tier.
- Native AWS: IAM, VPC and the full OpenSearch plugin set.
Pick WarpSearch for apps and catalogs, where you pay per request instead of keeping OCUs awake. For large log retention, OpenSearch Serverless Classic is cheaper today.
Sources
- OpenSearch Service pricing
- Scale to zero (NextGen)
- Serverless scaling and limits
- NextGen launch post
- WarpSearch pricing
NextGen hot storage at $0.30 per GB-month comes from the AWS price list entry for hot storage; AWS does not label it NextGen on the pricing page. Check the AWS Pricing Calculator. 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.