Skip to main content

Hetzner CX23

Helsinki DC Park 1

Benchmark complete
Finished Oct 16, 2026
Geekbench 7 single-core
597
Faster than 9% of 2 vCPU plans
Geekbench 7 multi-core
1,104
Faster than 6% of 2 vCPU plans
Disk 4K IOPS
40,269
fio random read and write, 50/50
Monthly price at test time
$6.49
170 multi-core points per dollar
Same price today

Server

  • Provider
    Hetzner
  • Plan
    CX23
  • Location
    Helsinki DC Park 1
  • vCPU
    2
  • RAM
    4 GB
  • Storage
    40 GB NVMe
  • Network
    Unknown
  • OS image
    ubuntu-24.04
  • Trigger
    Scheduled
  • Queued at
    Oct 16, 2026, 04:54:00 UTC
Visit Hetzner

Latest runs of this plan

Full history of this plan

CPU

Version Single-core Multi-core
Geekbench 7 597 1,104 Geekbench Browser

Scores from different Geekbench versions aren't comparable with each other. Charts only use Geekbench 7.

Geekbench 7 workloads

Workload Single-core Multi-core
File Compression
676
173rd of 191
972
130th of 135
Navigation
713
180th of 191
Single-core only
HTML5 Browser
563
177th of 191
Single-core only
PDF Viewer
636
181st of 191
Single-core only
Photo Library
511
179th of 191
1,033
125th of 135
Clang
596
175th of 191
1,164
126th of 135
Text Processing
446
182nd of 191
809
134th of 135
Asset Compression
565
185th of 191
1,078
130th of 135
HDR
634
171st of 191
1,242
110th of 135
Photo Editor
755
151st of 191
1,550
105th of 135
Ray Tracer
555
182nd of 191
1,124
130th of 135
Structure from Motion
682
177th of 191
Single-core only
Game Physics
489
183rd of 191
Single-core only
Video Encoder
553
179th of 191
Single-core only
Audio Encoder
696
173rd of 191
Single-core only
Video Player
573
177th of 191
Single-core only

Rank among the 191 runs with Geekbench 7 workload scores. Bars are relative to the best run. Geekbench doesn't run every workload multi-core, so some rows only have a single-core score.

Memory, TLS and scheduler

Memory bandwidth (STREAM)

Test Result Runs Spread
Copy higher is better 20 GB/s 3 10.4%
Scale higher is better 14.5 GB/s 3 20.5%
Add higher is better 16.8 GB/s 3 21.4%
Triad higher is better 16.4 GB/s 3 17.8%

Average of 3 runs. Spread is the standard deviation as a share of the average.

TLS and crypto (OpenSSL)

Test Result Runs Spread
RSA 4096 signs higher is better 322/s 3 8.88%
RSA 4096 verifies higher is better 21,358/s 3 3.09%
SHA-256 higher is better 303 MB/s 3 14.7%
AES-256-GCM higher is better 4.52 GB/s 3 6.68%
ChaCha20-Poly1305 higher is better 2.86 GB/s 3 3.28%

Average of 3 runs. Spread is the standard deviation as a share of the average.

Scheduler latency (schbench)

Test Result Runs Spread
Request latency, median lower is better 5.94 ms 3 8.02%
Requests per second, 90th percentile higher is better 325/s 3 1.96%
Request latency, 99.9th percentile lower is better 16.9 ms 3 2.85%
Requests per second, average higher is better 301/s 3 3.06%

Average of 3 runs. Spread is the standard deviation as a share of the average.

Disk

Block size Read Write Total
4K 78.5 MB/s (20,108 IOPS) 78.8 MB/s (20,161 IOPS) 157.3 MB/s (40,269 IOPS)
64K 859.7 MB/s (13,755 IOPS) 864.2 MB/s (13,827 IOPS) 1.68 GB/s (27,582 IOPS)
512K 1 GB/s (2,052 IOPS) 1.06 GB/s (2,161 IOPS) 2.06 GB/s (4,213 IOPS)
1M 1.07 GB/s (1,099 IOPS) 1.15 GB/s (1,173 IOPS) 2.22 GB/s (2,272 IOPS)

fio random read and write, 50/50, run by YABS against the boot disk.

Network

IPv4 transfer speed

Loading...
Target Protocol Send Receive
Clouvider, London, UK (10G) IPv4 5.28 Gbps 1.5 Gbps
Eranium, Amsterdam, NL (100G) IPv4 6.02 Gbps 2.48 Gbps
Leaseweb, Singapore, SG (10G) IPv4 1.13 Gbps 1.23 Gbps
Clouvider, Los Angeles, CA, US (10G) IPv4 1.06 Gbps 665 Mbps
Leaseweb, NYC, NY, US (10G) IPv4 2.21 Gbps 2.31 Gbps

Public iperf3 servers vary in capacity and availability. Compare a target only with itself over time.

Datastores

PostgreSQL

fsync speed (pg_test_fsync)

Method Operations per second Time per operation
One 8 kB write
open_datasync 2,873 0.348 ms
fdatasync 2,702 0.37 ms
fsync 1,213 0.825 ms
fsync_writethrough Not supported here n/a
open_sync 1,357 0.737 ms
Two 8 kB writes
open_datasync 1,457 0.686 ms
fdatasync 2,743 0.365 ms
fsync 1,170 0.855 ms
fsync_writethrough Not supported here n/a
open_sync 719 1.39 ms
open_sync at different write sizes
1 * 16kB open_sync write 1,320 0.758 ms
2 * 8kB open_sync writes 670 1.49 ms
4 * 4kB open_sync writes 354 2.83 ms
8 * 2kB open_sync writes 151 6.61 ms
16 * 1kB open_sync writes 76.9 13 ms
fsync on a non-write descriptor
write, fsync, close 1,163 0.86 ms
write, close, fsync 1,223 0.818 ms
Non-synced 8 kB writes
write 373,568 0.003 ms

pg_test_fsync against the Docker volume Postgres keeps its data on. fdatasync is Linux's default wal_sync_method.

pgbench at scale 10

Workload Clients Transactions/s Median p95 p99 p99.9 Failed
Read/write (TPC-B like) 1 669 1.37 ms 2.36 ms 3.11 ms 6.39 ms 0
Read/write (TPC-B like) 8 2,117 3.45 ms 6.54 ms 9.32 ms 14.2 ms 0
Read/write (TPC-B like) 32 2,114 11.7 ms 36.4 ms 58.8 ms 98.8 ms 0
Read only 1 8,501 0.104 ms 0.195 ms 0.287 ms 1.01 ms 0
Read only 8 23,546 0.311 ms 0.539 ms 0.688 ms 2.01 ms 0
Read only 32 22,913 1.19 ms 2.34 ms 3.14 ms 5.5 ms 0

Scale 10 (1,000,000 accounts), loaded in 2.73 s. 60 s per run, prepared statements. Latency comes from sampled transaction logs, and p99.9 is left out when fewer than 1,000 were sampled.

pgbench at scale 100

Workload Clients Transactions/s Median p95 p99 p99.9 Failed
Read/write (TPC-B like) 1 674 1.36 ms 2.31 ms 3.77 ms 6.39 ms 0
Read/write (TPC-B like) 8 2,080 3.59 ms 6.4 ms 9.48 ms 13.4 ms 0
Read/write (TPC-B like) 32 2,328 12.8 ms 22.4 ms 29.4 ms 38.8 ms 0
Read only 1 5,106 0.155 ms 0.403 ms 0.859 ms 2.22 ms 0
Read only 8 19,918 0.363 ms 0.673 ms 1 ms 2.41 ms 0
Read only 32 18,682 1.47 ms 3.09 ms 4.15 ms 6.87 ms 0

Scale 100 (10,000,000 accounts), loaded in 26.1 s. 60 s per run, prepared statements. Latency comes from sampled transaction logs, and p99.9 is left out when fewer than 1,000 were sampled.

MySQL

sysbench OLTP

Workload Threads Transactions/s Queries/s Median p95 p99 Errors
Read/write 1 133 2,660 6.91 ms 11.4 ms 18 ms 0
Read/write 8 434 8,688 17.6 ms 29.2 ms 38.2 ms 0
Read/write 32 449 8,975 69.3 ms 119 ms 148 ms 0
Read only 1 310 4,959 3.02 ms 5.57 ms 7.17 ms 0
Read only 8 850 13,601 8.9 ms 14.7 ms 18 ms 0
Read only 32 747 11,950 41.9 ms 61.1 ms 71.8 ms 0

10 tables of 100,000 rows, loaded in 17.7 s. 60 s per run, one connection per thread.

Redis

memtier

Mix Operations/s Hits/s Median p99 p99.9
1:10 set:get 37,293 33,902 1.54 ms 4.42 ms 7.33 ms
1:1 set:get 36,864 18,432 1.54 ms 4.54 ms 7.36 ms

1,000,000 keys of 32 bytes. 64 connections over 2 threads, 60 s per ratio.

Web frameworks

Framework Fortunes req/s Single query req/s JSON req/s Fortunes p99 at 256 connections
Next.js nextjs 297 1,068 1,595 1.71 s
Express express-postgresjs 4,440 6,448 16,473 318 ms
Fastify fastify-postgres 4,285 6,743 19,961 173 ms
NestJS nestjs 2,194 2,421 9,282 225 ms
Nuxt nuxt 3,685 5,479 13,932 543 ms
FastAPI fastapi 1,635 2,142 8,345 718 ms
Django django-postgresql 782 959 2,221 631 ms
Rails rails 1,318 1,986 7,091 270 ms
Laravel laravel 348 338 661 856 ms
Symfony symfony 430 463 1,466 700 ms

Peak requests per second across the 64 and 256 connection levels, measured with wrk. Each test has to pass the TechEmpower response checks first, or it's marked invalid.

Web apps

WooCommerce

Page mix

Step Requests/s Median p95 p99 Error rate
8 virtual users
Home page 0.962 1.65 s 1.88 s 1.94 s 0%
Shop page 0.962 2.18 s 2.65 s 2.68 s 0%
Category page 0.962 1.41 s 2.23 s 2.56 s 0%
Product page 0.962 1.47 s 1.81 s 1.85 s 0%
Product search 0.962 1.28 s 1.58 s 1.71 s 0%
32 virtual users
Home page 0.935 5.3 s 6.95 s 7.11 s 0%
Shop page 0.935 7.55 s 8.86 s 9.06 s 0%
Category page 0.935 7.03 s 8.59 s 9 s 0%
Product page 0.935 6.39 s 7.11 s 7.2 s 0%
Product search 0.935 5.8 s 6.83 s 6.95 s 0%

Cart and checkout

Step Requests/s Median p95 p99 Error rate
8 virtual users
Add to cart 1.58 2.08 s 2.34 s 2.53 s 0%
Cart 1.58 1.64 s 2.04 s 2.14 s 0%
Checkout 1.58 1.21 s 1.55 s 1.59 s 0%
32 virtual users
Add to cart 1.79 5.98 s 7.36 s 7.74 s 0%
Cart 1.79 6.49 s 7.2 s 7.48 s 0%
Checkout 1.53 5.4 s 6.21 s 6.38 s 0%

k6 with a fixed number of virtual users, 63 to 71 s per level. The error rate counts responses that weren't a 200.

Discourse

Signed in

Step Requests/s Median p95 p99 Error rate
1 virtual user
Topic page 0.746 582 ms 871 ms 1.13 s 0%
Topic JSON 0.389 296 ms 334 ms 335 ms 0%
Home page 0.389 516 ms 640 ms 679 ms 0%
Categories page 0.195 599 ms 734 ms 736 ms 0%
User activity page 0.178 372 ms 452 ms 469 ms 0%
4 virtual users
Topic page 1.36 1.21 s 1.72 s 2.48 s 0%
Topic JSON 0.673 667 ms 809 ms 964 ms 0%
Home page 0.673 1.13 s 1.5 s 1.62 s 0%
Categories page 0.369 1.33 s 1.71 s 1.97 s 0%
User activity page 0.337 809 ms 1.74 s 1.86 s 0%
16 virtual users
Topic page 1.23 2.57 s 3.11 s 4.16 s 0%
Topic JSON 0.649 1.96 s 2.39 s 2.57 s 0%
Home page 0.633 2.37 s 2.98 s 3.2 s 0%
Categories page 0.301 2.65 s 3.06 s 3.07 s 0%
User activity page 0.301 2 s 2.45 s 2.5 s 0%

Anonymous

Step Requests/s Median p95 p99 Error rate
16 virtual users
Topic page 27.9 148 ms 401 ms 619 ms 0.118%
Topic JSON 14 146 ms 380 ms 740 ms 0.235%
Home page 14 148 ms 413 ms 833 ms 0%
Categories page 7.01 155 ms 454 ms 900 ms 0.235%
User activity page 7.01 496 ms 910 ms 1.17 s 0.235%

k6 with a fixed number of virtual users, 61 to 63 s per level, after a 23 s warmup. The error rate counts responses that weren't a 200. Discourse ran with 4 app workers and 1024 MB of Postgres shared buffers, what its installer picks for a 4 GB box. It added a 2 GB swapfile first, the way its installer does, and didn't need it. The container peaked at 2,189 MB.

Ghost

Readers

Step Requests/s Median p95 p99 Error rate
1 virtual user
Home page 3.36 90.3 ms 109 ms 125 ms 0%
Post page 8.39 43.8 ms 58.2 ms 67.2 ms 0%
Tag page 1.68 56.2 ms 73.3 ms 96.5 ms 0%
Author page 1.68 59.1 ms 73.6 ms 90.8 ms 0%
RSS feed 1.66 67 ms 85.6 ms 98.4 ms 0%
4 virtual users
Home page 4.32 253 ms 339 ms 443 ms 0%
Post page 10.8 167 ms 237 ms 267 ms 0%
Tag page 2.18 152 ms 215 ms 233 ms 0%
Author page 2.14 166 ms 223 ms 267 ms 0%
RSS feed 2.14 159 ms 238 ms 289 ms 0%
16 virtual users
Home page 4.43 936 ms 1.12 s 1.22 s 0%
Post page 11 754 ms 949 ms 1.06 s 0%
Tag page 2.21 497 ms 673 ms 744 ms 0%
Author page 2.21 516 ms 649 ms 676 ms 0%
RSS feed 2.2 521 ms 739 ms 822 ms 0%

Staff in Ghost Admin

Step Requests/s Median p95 p99 Error rate
1 virtual user
Post list 1.87 77.6 ms 98.3 ms 120 ms 0%
Post editor 5.65 31.6 ms 44.5 ms 53.8 ms 0%
Saving a draft 5.65 79 ms 99.3 ms 114 ms 0%
Member list 1.87 95.5 ms 133 ms 149 ms 0%
4 virtual users
Post list 2.4 106 ms 143 ms 161 ms 0%
Post editor 7.31 52.3 ms 76.4 ms 93.3 ms 0%
Saving a draft 7.34 398 ms 521 ms 588 ms 0%
Member list 2.45 117 ms 154 ms 246 ms 0%

k6 with a fixed number of virtual users, 60 to 61 s per level, after a 20 s warmup. The error rate counts responses that weren't a 200. Ghost ran the way its ghost-docker setup runs it: one Node process, and MySQL with its stock 128 MB buffer pool. Memory peaked with Ghost at 285 MB and MySQL at 619 MB.

Wagtail

Visitors

Step Requests/s Median p95 p99 Error rate
1 virtual user
Home page 0.594 219 ms 246 ms 262 ms 0%
Blog index 0.594 794 ms 862 ms 863 ms 0%
Blog post 1.75 70.3 ms 98.1 ms 116 ms 0%
Bread page 1.76 77.7 ms 93 ms 97.2 ms 0%
Location page 0.594 72.6 ms 83.5 ms 89.5 ms 0%
Site search 0.577 123 ms 205 ms 220 ms 0%
4 virtual users
Home page 1.13 498 ms 835 ms 903 ms 0%
Blog index 1.16 1.49 s 2.4 s 2.6 s 0%
Blog post 3.39 142 ms 249 ms 327 ms 0%
Bread page 3.44 165 ms 245 ms 301 ms 0%
Location page 1.18 142 ms 225 ms 258 ms 0%
Site search 1.11 223 ms 461 ms 708 ms 0%
16 virtual users
Home page 1.33 1.36 s 1.72 s 1.89 s 0%
Blog index 1.39 3.47 s 3.91 s 4.09 s 0%
Blog post 4 788 ms 1.17 s 1.27 s 0%
Bread page 4.08 805 ms 1.14 s 1.28 s 0%
Location page 1.38 830 ms 1.09 s 1.27 s 0%
Site search 1.31 954 ms 1.35 s 1.49 s 0%

Editors in the Wagtail admin

Step Requests/s Median p95 p99 Error rate
1 virtual user
Admin dashboard 0.243 228 ms 318 ms 405 ms 0%
Page explorer 0.519 892 ms 1.08 s 1.17 s 0%
Page editor 0.779 402 ms 531 ms 633 ms 0%
Image library 0.26 161 ms 621 ms 1.64 s 0%
Admin page search 0.243 95 ms 221 ms 245 ms 0%
4 virtual users
Admin dashboard 0.417 504 ms 1.01 s 1.1 s 0%
Page explorer 0.867 2.03 s 3.66 s 3.81 s 0%
Page editor 1.32 949 ms 1.73 s 1.8 s 0%
Image library 0.433 399 ms 1.07 s 1.33 s 0%
Admin page search 0.417 244 ms 765 ms 837 ms 0%

k6 with a fixed number of virtual users, 61 to 62 s per level, after a 21 s warmup. The error rate counts responses that weren't a 200. Wagtail ran the way bakerydemo ships it: uWSGI with 2 workers of 4 threads each, whatever the box, and Postgres and Redis on stock settings. Memory peaked with Wagtail at 289 MB, Postgres at 155 MB, and Redis at 26 MB.

Advertised vs. measured

Advertised Measured
vCPU 2 2
CPU Shared vCPU 0% of busy CPU time stolen
RAM 4 GB 3.7 GiB
Storage 40 GB NVMe 37.5 GiB
Architecture x86_64 x86_64
Virtualization KVM KVM

Guests usually report slightly less RAM and disk than advertised because of kernel reservations and filesystem overhead.

Detected hardware

CPU model
Intel Xeon Processor (Skylake, IBRS, no TSX)
CPU cores
2
CPU frequency
2095 MHz
AES-NI
Yes
VM-x/AMD-V
No
RAM reported
3.7 GiB
Swap reported
0 MiB
Disk reported
37.5 GiB
Virtualization reported
KVM
CPU steal
0% of busy CPU time stolen
Distro
Ubuntu 24.04.4 LTS
Kernel
6.8.0-138-generic
Architecture
x86_64
IPv4 reachable
Yes
IPv6 reachable
Yes

Advertised at test time

Plan
CX23
Product line
cost_optimized
Location
Helsinki DC Park 1
vCPU
2 shared vCPU
Architecture
x86_64
RAM
4 GB
Storage
40 GB NVMe
Virtualization
KVM
Network
Unknown
Transfer
21.99 TB
Port 25
On request
Monthly price at test time
$6.49
Hourly price at test time
$0.010
YABS version
v2026-09-20
Ran at
Oct 16, 2026, 04:55:07 UTC
Parser version
5

Run timeline

Queued
Oct 16, 2026, 04:54:00 UTC
Server created
Oct 16, 2026, 04:54:55 UTC
First update from the server
Oct 16, 2026, 04:54:55 UTC
Completed
Oct 16, 2026, 06:51:12 UTC
Server deleted
Oct 16, 2026, 06:51:12 UTC
Cost of this run
$0.020
Watch party 0 watching

This party closed at . It's read only now.

Watch party

Loading the party