As you said, you don't send different types to do different jobs in sequence. Even after you had achieved air dominance, your strike packages will never be going in alone and unsupported. At all times, you will have the full spectrum of assets in deployment or on standby.
They enemy could hide his SAMs until day 60 or 600, it doesn't matter, there will always be SEAD and DEAD assets on station to take them out as and when they reveal themselves or are discovered.
Sure, you cannot plan ahead and mark on your calendar, but at some point, your assessment would be that they enemy's air force and air defences have been degraded so much that you no longer need first day of war level of stealth.
I would expect the JSF to spend the vast majority of its life hauling bombs on external pylons, and when they do that, they effectively give up pretty much all the stealth advantages their costly and limiting stealth design gives them.
You get all the costs but little of the benefit.
The argument against going full stealth is that it is better to just buy conventional fighters in greater numbers and/or at lower cost then getting stealths only to hand bombs on their wings.
The whole point is that if the "low" are conventional fighters, you save so much you can afford enough of your "hi" end asset you could not only overwhelm opfor's far smaller top end fighters, you could be able to use your top fighters against their "low" end fighters such that there is little chance any of their "low" end fighters would survive to even had a go at your own "low" end fighters.
We have to also remember that 5th gens only hold a distinct advantage in BVR against even 4th gen fighters. In a furball, the distinction between 4th and 5th gen pretty much melts away to near nothing, with the possible exception of the F35, who may really struggle against 4th or even 3rd gens in a dogfight because of the compromises needed for stealth and its 3 in 1 design.
If a mixed force of 10 J20s and 25 J10s came up against an opposing force of 5 F22s and 15 F35s, the J20s only need to keep the opfor occupied long enough for the J10s to get within WVR and the numbers alone would probably be enough overwhelm the opfor even if we are extremely generous and assume the F35s can hold their own in WVR against a J10.
Stealths operating with conventional assets doesn't mean they have to all fly formation.
Just because they are operating in the same airspace does not mean you can deduce the position of enemy stealths based on where their conventional assets are operating.
Moreover, operating with conventional fighters gives you options not available to an all stealth force.
I am going to assume J20s, F22s and PAK FAs will be able to quite effectively assassinate opfor AWACS or at least pose such a grave threat as to force them to withdraw so far back as to no longer be able to provide meaningful support until the enemy 5th gen threat could be neutralised.
Thus, if we rule out AWACS and ground based radar support, and have it purely as a fighter vs fighter game, the side with conventional fighters would actually hold the advantage.
You do not want your stealths to be using their radars actively scanning, even in LPI mod, as it would be far easier for opposing 5th gens to detect enemy active radar scanning and use that to get a locational reference for opposing 5th gens than it would be to pick up enemy 5th gens with your active scans.
Your conventional fighters wouldn't really care, since the enemy would likely be able to detect them even if they remained silent.
That actually frees up the conventional fighters to go and scan at full power and act like hounds to flush out the enemy 5th gens.
The enemy is left with the difficult choice of risking being detected by the enemy 4th gens' active scans as they continue to try and find the opfor 5th element, or engaging the enemy 4th gens and probably exposing themselves to attack from opfor 5th gens.