There really is no "rule" that works. The size of tank that any fish requires is a matter of the fish's length, its mass (thicker fish opposed to thin fish), how many of its species may be needed, its behaviours (some need room to actively swim, others less), and the other species in the tank. This takes a bit of explaining, so I will expand on any of these if asked.
I had a male/female pair of Pantodon buchholzi for many years back in the 1980's. They are best on their own, in a shallow tank, 24-30 inches length. They spend most of their time motionless, except when feeding or when interacting. They like to hover among floating plants. Drop in an insect and they spring into action. They will eat, or attempt to eat, any fish that comes to the surface.
For anyone interested in this very unique fish, some info from a profile I authored a few years back:
This fish was described by W.C.H. Peters in 1876. It is the only species in the genus, and there are no other genera in the family. The name Pantodontidae is derived from the Greek pan [= all] and odous [= tooth, teeth]. The species epithet honours a Professor Buchholz who discovered this species. Two sub-species described by C. Bruning in 1911 were determined by Gosse (1984) to be conspecific with the subject species.
Pantodontidae is closely related to the family Osteoglossidae (Arowana) and these along with five other families that include the Arapaima, Elephantnoses and Knifefishes, belong to the Order Osteoglossiformes. The name derives from the Greek osteon [= bone] and glossa [= tongue] plus the Latin forma [= shape], and in English these fishes are referred to as the bony tongues. This is a very primitive order, with fossil records as far back as the late Jurassic period (roughly 161 to 145 million years ago). To put this into perspective, this was the period when the supercontinent Pangaea broke up into the two supercontinents Laurasia and Gondwana, which in time further divided into the continents as we know them today.
There are two distinct populations known in the wild, one in the Niger River basin and one in the basin of the Congo River, and only recently have these been shown to have significant genetic differences that have existed for at least 57.2 million years. In spite of this, the species is an example of what is termed "morphological stasis," which means that the morphology (the form or shape) has remained virtually unchanged throughout the species' existance. This constancy is believed to be greater than that of all vertebrate examples in fossil records (Lavoue, et al, 2011; cited in Dawes, 2011). Further study may determine that these are in fact two distinct species; the genetic differences between the two populations are so wide that "they might even be reproductively isolated owing to genetic incompatibilities." (Dawes, 2011).
Byron.