I think the main point that is missing is that the standard model is very close to symmetric, so if you have some process that generates systems of 4 or 8 or however many particles from a neutral boson then exactly the same process should occur producing all the anti-particles of the above e.g. if you think you make
1 proton, 1 neutron, 1 electron and 1 neutrino
in such a way that all the numbers balance then you should also see a process that makes
1 anti-proton, 1 anti-neutron, 1 anti-electron and 1 anti-neutrino
Its also worth pointing out that pair production is not specifically an electromagnetic thing. For example the Z0 boson pair produces fermion+anti-fermion pairs in a very similar way to the photon and W bosons decay to lepton+anti-neutrino pairs in a similar way (although W bosons are charged so its a little different). The Higgs boson also decays into quark-anti-quark,lepton-anti-lepton or even boson-anti-boson pairs.
I did not claim that this is how matter was generated in the Big Bang, especially because for now there exists no theory for such a process.
My point is that this is a strange coincidence as it would provide matter in the right proportions. Unlike for electromagnetic interactions, there might be some asymmetry making the process generating the particles more probable than the process generating the antiparticles. Because the 2 processes are decoupled, they might have happened at different rates, which is not possible with electromagnetic generation and annihilation.
Of course, the correct explanation for Big Bang might be completely different, only the fact that there was no simultaneous generation of matter and anti-matter is certain.
The neutral Z0 boson behaves like the photons, so I did not mention it.
The charged W bosons cannot generate particle/anti-particle pairs because that would violate the conservation laws. They generate simultaneously 4 elementary particles: quark, anti-quark, lepton and anti-lepton.
The quark and the anti-quark remain bound as one meson.
The reverse event is also possible, but highly improbable, because it is unlikely for a meson, a charged lepton and a neutrino to collide simultaneously, in order to annihilate into W bosons. What happens frequently is the equivalent event when 2 of the 3 collide and transform into a W boson together with the antiparticle of the third.
These type of events involving W bosons are what I was referring to as the generation/annihilation of systems of 4 elementary particles through weak interactions.
W^+ -> positron + electron neutrino
W^- -> muon + muon antineutrino
are allowed as are processes like
W^+ -> up quark + anti-down quark
Unlike the stable photon, the W bosons have a negligible lifetime.
So in fact those processes listed by you that result in 2 particles are 4-particle processes.
Two particles collide, either a quark and an anti-quark giving other 2 particles, a lepton + an anti-lepton (in your exemples positron + electron neutrino or muon + muon antineutrino) or a lepton collides with an anti-lepton giving a quark and an anti-quark (in your exemple up quark + anti-down quark).
So there are always 4 particles, either quark, anti-quark, lepton and anti-lepton, or there may be both at input and at output a lepton and an anti-lepton, which includes the case of scattering through weak interactions, or there maybe both at input and at output a quark and an anti-quark, which includes the case when a meson decays into another meson through a weak interaction.
All the interactions involving W bosons are 4-particle interactions, where the W boson is an intermediate particle that exists only during a negligible time.
The 4 particles may be all at the input, all at the output, 3 at the input and 1 at the output, 1 at the input and 3 at the output or the most frequent case, 2 at the input and 2 at the output.
Like all the interactions between elementary particles all these cases are equivalent and all the variants are interchangeable by moving a particle from the input to its anti-particle at the output or vice-versa.
In most cases you need to pay attention only to the 4 particles participating in the weak interaction and not to the intermediate W boson.
The existence of the W boson matters only when you need to compute various things because this intermediate particle transforms diagram nodes with 4 edges that cannot be computed into diagram nodes with 3 edges that can be computed with the methods of quantum electrodynamics.
There are many.
The Big Bang theory that is useful starts with matter at very high temperature and density, but otherwise not different from the matter that we know. That happens at some uncertain time interval after the beginning of Big Bang.
We may try to extrapolate towards time 0 and increasing temperatures and densities, but that reaches soon values of temperature and pressure high enough that we do not really know how matter behaves in those conditions and it does not matter anyway because it does not influence what happens later.
For the later evolution, it is enough to start the modeling of Big Bang from the moment when the temperature became low enough, e.g. of some tens of MeV, i.e. when matter was too hot for nuclei and atoms to exist, but cold enough so that it consisted of a plasma composed of protons, neutrons, electrons, positrons, photons and neutrinos, with only negligible quantities of heavier particles.
For the time 0 there is no theory that can predict anything quantitative.
2 particles, one of which is the anti-particle of the other form a system where the sum of all quantities that are conserved can be 0, so a transformation generating or destroying both particles is possible.
Nevertheless there are also other systems of particles where the sum of all conserved quantities can be 0 so there should exist transformations that generate all of them simultaneously.
Besides the systems of 2 particles, where generation and annihilation is well known, there are systems of 4 particles with zero sum of all conserved quantities (e.g. a quark, an anti-quark, an electron and an anti-neutrino) and also systems of 8 particles e.g. the 3 kinds of u quarks + the 3 kinds of d-quarks + electron + anti-neutrino.
The systems of 4 particles can actually be generated and destroyed in weak interactions, similarly like the the systems of 2 particles can be generated and destroyed in electromagnetic interactions.
For the systems of 8 particles, there is currently no theory about a mechanism of generation or annihilation, but my bet is that this is how matter was created during Big Bang (those 8 particles aggregate into 1 proton, 1 neutron, 1 electron and 1 neutrino).
For now we have a theory of Big Bang not from the starting point but from a little later when there already was normal matter as we know it.
There is no theory yet for the origin of Big Bang, but there is absolutely no reason to believe that at the beginning there was an electromagnetic interaction generating simultaneously equal numbers of particles and anti-particles. For that to be true, the matter and anti-matter should have been preceded by electromagnetic waves i.e. photons, of equivalent energy with the mass of the generated matter and anti-matter, which would then annihilate restoring the previous photons. Such a theory goes nowhere, so it does not match reality. Whatever started Big Bang, it was not an electromagnetic field generating pairs of particles and anti-particles.