For a relatively optimized party, there isn't a level where it is balanced (or even worth using).
The optimal situation to use Elemental Bane is when you have a group of people using elemental attacks against a single target, the attacks will continue for some time (aka the target has a lot of hp, or the attacks do very little damage on their own), the target has Resistance to that elemental type, and that target fails the Constitution saving throw.
If a target has a low constitution saving throw, it typically will not have a huge amount of hp. Enough hp to last dozens of attacks and a good con save are often correlated.
D&D parties typically don't do 'very little' damage with their attacks (when they can help it), and often don't do the same damage type - the barbarian does Slashing, not Fire, again, usually.
Single enemies fighting against a 7th level D&D party often have very good saves, legendary resistances, or if they lack these traits are not tough encounters that a 7th level spellcaster should use a 4th level spell slot to defeat.
An optimized party will generally try to have access to multiple damage types, and when facing against an opponent roll Arcana, Nature, or Religion to try to recall if it resists fire or is weak to sunlight or what have you. Likewise many DMs allow players to work out if a target is resistant or immune to a type of damage by that foe's reaction to eg. being shot with flames (if they don't get burned that's a good clue).
Ergo outside people blithely applying it to things like 'maximum damage challenges', Elemental Bane is not going to be optimal for the majority of situations that a D&D party finds themselves in. A situation where a 7th level spellcaster and a gaggle of firebolt using apprentices face off against a fire-resistant demon would be a great situation for Elemental Bane, but that is not the typical D&D party composition.
A less optimized (Bob the Sorcerer only ever chooses Fire spells) and caster-heavy dnd party, especially a larger one (6+ players) where the DM has artificially inflated monster HP without changing other stats, and treats fights as slugfests (intelligent enemies do not use traps, tactics, magic items or spells but instead wade into melee and attack at random or only hit the full plate + shield tank over and over again etc), with a party that either is outright told that the creature is resistant as opposed to immune or metagames, will potentially see Elemental Bane outperform other spells - likewise in a situation where the party has multiple elemental-damage using allies (summoned via Planar Binding or other methods) it will do a lot better.
(Situations with high-hp but otherwise unimpressive stats enemies and unoptimized parties are more common than you'd think in 5e - often hp was inflated as a method to make fights 'last longer', or at least that's what I assume based on complaints about fight length on various forums and the suggested solutions to make them last longer. That advice and complaint has been tapering off considerably, and now the opinion on various forums seems to weigh heavily against this form of length-padding meaning that it is likely less common than in the early years of 5e's existence (again, judging by forum posts on the topic). But it probably still exists especially in DMs who started DMing in early 5e era with their style informed by the mode of that time period.)
However even in this ideal environment, elemental bane is not necessarily better than a summon spell upcast to 4th level, or a spell like Hold Person or Fear which has a very meaningful effect on a failed save, in some cases effectively defeating the enemy. Even with a situation of high hp low stats enemy, unoptimized party who all use the same elemental damage type etc, it could still be worse than another spell cast out of a 4th level slot.
Thus, overall Elemental Bane is a very weak spell. In its ideal environment, it's still a maybe. In a typical environment, it's very rarely going to be a good idea. In an optimized party that uses common tactics like 'not casting fire spells against the fire resistant enemy', it is unlikely to be a good idea to use.
Even if it was a 1st level spell, the Action cost is likely better spent elsewhere. Having a chance to add a bit of extra damage and ignore a resistance is just too situational, it relies on a party with the right elemental damage type, and it relies on knowing the creature is resistant (and not say, Immune).
The level at which it is best balanced is fairly unknowable. The party of 8 characters fighting a 167hp guy who resists everything and does 1d8+4 damage is probably a thing and in that milieu a 2nd level spell that gets rid of a resist and adds 2d6 damage to their 1d8 cantrips is probably pretty great. In a more optimized group fighting lower-hp, higher-saved, and much higher damage foes, a 2nd level spell that does that is mostly going to be a waste of a turn.
It unfortunately interacts with 'how optimized is the party' and 'how fight-padding is your DM' and those are big deals for 5e and very variable matters of tone and table style which will hugely inform the balance point of the spell.