That's not entirely correct. There is no specific level to which uranium (or any other fissile material) must be enriched before supercriticality may occur, the level of enrichment simply determines the critical mass required. For instance, the critical mass for pure U-235 is around 50kg, while for 20% enriched U-235 you'd need about half a ton or so. [1]
In the case of Chernobyl, the concern was that upon contact with water (which acts as a neutron moderator in these circumstances) and given the inevitable sequence of steam explosions that would follow, some regions of superheated corium might be forced into prompt criticality events. These would technically qualify as nuclear explosions, but the sequence of events required to get from this set of increasingly improbable assumptions to anything remotely close to a megaton range explosion is, as I said, improbable to the point of being trivially dismissable.
[1] EDIT: I realized I should probably point out for clarity that I'm referencing the traditional measures for a critical mass - the mass required for a homogenous sphere of a given material to go critical. Determining the critical mass for a highly inhomogenous material of complex geometry, highly varied temperature and potentially surrounded by neutron reflecting substances is not a trivial task.
Prompt critical is a necessary, but not sufficient condition for a nuclear weapon to go boom. In a weapon the neutron multiplication time is a million times faster than in a prompt criticality transient in a thermal reactor. Otherwise it would disintegrate before it could develop any significant yield.
But yes, certainly the melted fuel becoming critical and/or causing more steam explosions could have made the Chernobyl accident worse than it already was.
In the case of Chernobyl, the concern was that upon contact with water (which acts as a neutron moderator in these circumstances) and given the inevitable sequence of steam explosions that would follow, some regions of superheated corium might be forced into prompt criticality events. These would technically qualify as nuclear explosions, but the sequence of events required to get from this set of increasingly improbable assumptions to anything remotely close to a megaton range explosion is, as I said, improbable to the point of being trivially dismissable.
[1] EDIT: I realized I should probably point out for clarity that I'm referencing the traditional measures for a critical mass - the mass required for a homogenous sphere of a given material to go critical. Determining the critical mass for a highly inhomogenous material of complex geometry, highly varied temperature and potentially surrounded by neutron reflecting substances is not a trivial task.