Blog speed test: how fast is this Khalid blog?

We ran a blog speed test on this blog on 3 November 2026 to check Khalid's speed claims against real pages. The short version: Khalid answers in about half a millisecond on the server, and most of a reader's wait is the trip across the internet. Below are our own numbers, how we measured them, and a file of the raw data.

How fast does Khalid answer? #

Khalid answered each page in about half a millisecond on the server in our blog speed test. That is the time from a request reaching the program to the first byte of the reply. It is fast because Khalid builds each page once, when you publish, and then serves the finished page from memory.

Server times in our blog speed test #

We sent twenty requests to each of seven pages from the server itself, so no network sat in the way. The median for every page came in under six tenths of a millisecond, as the chart shows.

Server response time by page, median of twenty requests

Page	Milliseconds
Home	0.56ms
AI answers guide	0.59ms
Answer capsule	0.46ms
Topical map explained	0.52ms
Building a topical map	0.56ms
Topical map links	0.54ms
How this blog works	0.42ms

For comparison, Khalid's own release checks require the first page in under a tenth of a second, and a post page to render in under one millisecond at the median (Khalid on GitHub). Our pages stayed well inside both.

How long does a reader wait? #

A reader in our blog speed test waited about two tenths of a second for the first byte of a page over the public internet. That time includes the journey to the server and back through Cloudflare. So almost all of the wait is distance, and only about half a millisecond of it is Khalid.

What a reader waits for #

We also sent ten requests to each page from our own computer over the public internet, on the same day. Every median came in at about a fifth of a second. A reader closer to the server in Germany would wait less, and one further away would wait more.

$$\text{Share}=\frac{\text{server time}}{\text{public time}}\times 100$$

With our medians, that is about half a millisecond divided by about two hundred milliseconds, times a hundred. So Khalid takes up about a quarter of one percent of the wait.

How heavy are the pages in the blog speed test? #

The pages in our blog speed test weighed between about a hundred and about three hundred and fifteen kilobytes in all, as Khalid's own audit counts them. The HTML of each page, the part search engines read first, stayed under forty kilobytes. Most of the rest is the cover picture, served in sizes that fit the reader's screen.

How much each page weighs #

Total page weight by page, from Khalid's audit

Page	Kilobytes
Home	314.4KB
AI answers guide	272.5KB
Answer capsule	211.0KB
Topical map explained	279.6KB
Building a topical map	258.8KB
Topical map links	257.0KB
How this blog works	104.1KB

Page weight matters for search too. Steve Toth reported in February 2026 that Googlebot reads only the first 2 megabytes of a page (his post, 2026). Our heaviest page used about a sixth of that, so search engines read every word.

How did we run the blog speed test? #

We measured with plain command-line tools, so anyone can repeat the test. Server times came from twenty requests per page made on the server itself. Public times came from ten requests per page over the internet. Page weights came from Khalid's audit command, run on the same pages.

Download the raw numbers #

The file holds one row per page, with the median server time, the median public time, the HTML weight and the total weight. You may reuse it under a Creative Commons Attribution licence. Please name the SerpDad Blog and link this page.

Raw measurements as a CSV file

To see how a Khalid blog compares with WordPress on more than speed, read our Khalid and WordPress comparison. For every feature this blog uses, see how this blog works.