Gendustry is a particular thing. It allows to mass produce bees, in the same way you mass produce iron by throwing ore it into macerator.
Without that ability to duplicate - you would have to work much harder to get several bees of wanted resource. Almost impossibly hard. That is where it becomes close to adequate. And even still - for things to be balanced, they should be balanced regarding Gregtech. By that I mean that production costs of things with bees should be comparable to that of in the Replicator. Bees might go as alternative, they may go as a cheaper alternative (as you need to climb up) but production costs\processing of EACH item should be comparable to that of a replicator.
Here's my experience as someone who has bred every single bee available in the game to date: Gendustry isn't necessary for me to duplicate bees at all. I duplicate my princesses the old fashioned way, by putting a PRISTINE princess in an apiary with 3 necrotic frames (or 1 oblivion) and a stack of drones. I automate the apiary and let it go until the princess takes on all of the traits of the drones.
Sure, I certainly can collect liquid DNS and liquid protein and combine them with a complete genetic template to create princesses, but those princesses will be ignoble every single time. I have no use for ignoble bees at all other than to trade to villagers. Ignoble bees don't last more than 1 full day (RL) playing time. Anyone running production bees that understands how bees work will run only pristine princesses/queens. And the only way to get pristine princesses to apply traits to is to go out in the world and collect them. In the past month I have broken thousands and thousands of hives in the wild to get the amount of pristine bees that I need to ramp up my production to levels where I can sustain resource collection from it (and I'm still not there with over 50 alvearies and 10 industrial apiaries). I have spent significantly more time running around the world with a bee scoop looking for hives than I have core sampling for ore veins.
In my opinion the power of Gendustry lies in the ability to reliably and accurately mutate/breed bees without fail, and to a lesser degree reducing the amount of time it takes to breed preferred traits into bees. It saves time. That's why I use gendustry, and that's why I mined up many veins of uranium and turned it into Mutagen and also turned many diamond veins of diamonds into labware. But duplicating bees? Nope. Ignoble bees suck. I've never even once used the Genetic Replicator, even though I have one and could use it at any time. It's useless.
As for a solution, yeah I think Osmium Bees should go. There's also iridium bees which give 3 types of combs: Iridium, Tungsten, and Platinum. Tiny osmium dust is a rare chance byproduct of the iridium comb. Get rid of the osmium bees that give shards. In fact, though it pains me to say this most of the Magic Bees that give shards instead of combs are sort of broken and just way too easy to get product out of. Leave the iridium bee and anything else that makes a comb. Require the 3 combs from the iridium bee to be centrifuged in EV voltage or higher. Naquadah comb should be EV voltage or higher also, and so should Naquadah dust to Enriched Naquadah dust be EV+. This makes the highest end materials (tungsten, iridum, osmium, platinum, naquadah) require at least EV era equipment and also large amounts of EU in order to process.
Additionally I skipped right past the MK1 and MK2 naquadah reactors for the MK3 last night and now have 1.6 billion EU stored in just a few blocks with a very simple setup. Tier the MK3 reactor recipe with fusion as suggested as it's actually simpler to run a MK3 naquadah reactor than it is Mk1/2 because you can easily turn the enriched naquadah dust into cells, whereas with bolts you need to put it in a EBF, vacuum freeze it, and then extrude it. Plus the MK3 naquadah reactor is more efficient than 1/2, which is backwards from most GT power gen?
Edit: FWIW centrifuging Naquadah dust ( the dust, not the comb) to get enriched naquadah tiny dust is already a HV centrifuge recipe at 320EU and it's at a cost of just over 1million EU per dust. You've got a 50% chance to get enriched naquadah tiny dusts, so that's around 18million EU on average to get 1 enriched naqaudah dust. Still a net gain in EU, obviously, but it's far from free.