I've seen it quoted as being 13-25 million years, then some sources say it's only ten, some say six.
It's annoying to pin down a proper number.
But still, anything with six zeroes after it is NOT a short space of time.
The reason the number fluctuates is because the boundaries are sort of arbitrary. Like if you were on a rollercoaster, for example, would you call the beginning of your vertical ascent starting at a 10 degree angle? Or 20 degree? 80 degree? Etc.
Some ID advocates and critics, they want to use "instantaneous" as their select word, but if the Cambrian explosion were really instantaneous, or lets say less than 5 million years, then it would exclude 90% of phyla. If it were 10 million years, it might include 50%, if it were 20 million, it might include something like 80% etc. To really encompass organisms of the Cambrian explosion, 90+%, really it would have to be something like 30+ maybe even close to 40 million years long.
And part of it too is this lack of discussion around the fact that animals were soft bodied before the Cambrian in large part. So for example, arthropod diplichnites, foot tracks appear tens of millions of years before the first actual hard shelled body. And so it's not even that animals even "appeared", rather they lived beforehand and just evolved shells, giving this false appearance of an "explosions".
Which is why 50 years ago an explosion was a good way to describe it, but over time it's gradually being revealed to be more like something similar to other radiations observed throughout the fossil record.
And one book that describes this in a nice simple way is Donald Protheros story of life in 25 fossils. You can pick it up on Amazon for maybe just 20 bucks. He gives sourced research and is well credentialed in the field.
If people based the story of life only on what is observed in the fossil record, we would have phylum appearing all over the place, some in the Eocene some in the mid to late Paleozoic, some in the Mesozoic, some in the Precambrian, some in the Cambrian etc. Because fossils just aren't particularly precise in where they appear, some animals don't even have a fossil record and so based on the fossil record we might think that they don't even exist. So when we see all these fossils appearing in the Cambrian, we can't confuse the appearance of fossils with the question of if those animals actually did or did not exist. And we know this because we have soft bodied lagerstaaten that predate hard Cambrian shelled fossils. They're rare, But we know they're there. That and we have trace fossils like Rusofycus and diplichnites that also predate hard shelled fossils. But obviously the animals were alive, else their foot tracks wouldn't be there. But according to critics, they might argue that arthropods appeared in the Cambrian explosion, but that's actually wrong.
So the fossil records use doesn't work for critics of ID.
Alternatively, if we base our understanding on genetics, the problem is of course worse because all of these phylum (or more specifically superphylums or classes) actually appear well before the Cambrian in the precambrian. Which is also is equally a problem for critics. And this is problematic largely because most pre-cambrian rock is old and recycled and metamorphosed. And so digging up microscopic soft bodied animals and old recycled rock tends to be problematic. Which is why the whole topic is ridiculous when intelligent design advocates want to put so much focus on The Cambrian explosion.
I gave this analogy the other day but using the Cambrian explosion to judge the theory of evolution is like trying to use grade school video footage of Michael Jordan to judge how good of a basketball player he was in the course of his life and professional career.
If the best arguments against evolution are centered around soft-bodied microscopic organisms in billion-year-old tectonically recycled pre-cambrian rock, For practical purposes, they've already lost.