Though I see different estimates of the size of Oppau 1921; the material that exploded was under 500T of fertilizer (17% of Beirut), which means it should be under 250kT of TNT equivalent (I've seen AN assessed at 42% as explosive as TNT by mass). But based on crater size, it looks like estimates are more like 1-2 kT, similar to Beirut. Also, windows blown out "more than 25km away" is pretty similar in blast radius to Beirut (I've seen "15 miles" quoted, IIRC), suggesting a similar magnitude.
It's actually remarkable, comparing against other blasts of similar quantities of material or less, how few fatalities there were in Beirut, especially given that it was inside the city. Brest 1947 is larger with less fatalities, but they had towed it out to sea three hours before it finally exploded.
Though I think that, when it comes to non-nuclear explosions, https://en.wikipedia.org/wiki/Halifax_Explosion is still the title-holder. What a privilege.
Take a roll of fast burning cord and set it on fire when rolled out: a slight fizzing and some minor damage to plants. Wrap 10 turns of it around a tree, set it on fire and you'll have a cut cleaner than the sharpest axe cutting your tree clean in half.
The Halifax one is just terrible. I went to the memorial site when visiting the city just before it was removed. Incredible but true: the city lost the memorial sculpture.
Now if something like a tank were blown over, then I'd be impressed.
"The workers needed to use pickaxes to get it out, a problematic situation because they could not enter the silo and risk being buried in collapsing fertilizer. To ease their work, small charges of dynamite were used to loosen the mixture.
This seemingly suicidal procedure was in fact common practice. It was well known that ammonium nitrate was explosive, having been used extensively for this purpose during World War I, but tests conducted in 1919 had suggested that mixtures of ammonium sulfate and nitrate containing less than 60% nitrate would not explode. On these grounds, the material handled by the plant, nominally a 50/50 mixture, was considered stable enough to be stored in 50,000-tonne lots, more than ten times the amount involved in the disaster. Indeed, nothing extraordinary had happened during an estimated 20,000 firings, until the fateful explosion on September 21.
[1] https://en.wikipedia.org/wiki/Oppau_explosion