The thing is you don't even need to believe the whole gender is a social construct thing to see that this is bad teaching and bad science.
The teacher may well believe that there is accross the population some marginally greater frequency of the abilities, aptitude and interest needed for high level physics amongst boys than amongst girls. He may be right and he may be wrong. But it isn't necessarily the best argument to have with him.
Even if his observation that "girls are better at biology, and boys do better at the physics side of things'' is not 100% down to cultural conditioning, he shouldn't use that in any way to guide his teaching or assessment of individual pupils (and if he does it is likely to be a self-fulfilling prophesy).
I may seem like splitting hairs, or ceding ground....But if you go in there with an argument that starts with 'there is absolutely no chance of any difference between the distribution of innate abilities in physics amongst girls and boys', the teacher/head may silently agree to differ with you and try to palm you off with a mealy mouthed apology.
Where as if you go in there with the argument that each child should be judged against their own potential, previous performance, performance in related subjects like maths etc...not by a broad stereotype - then you are keeping the discussion on the principle that really matters here.
My advice would be to think about what you want before you go in, on a practical level, otherwise they will just give you an apology without looking to change anything.
If you just go in and say you are angry/upset, the head will see the problem as you being angry and try to palm you off with something to calm you down (which would solve the problem as s/he sees it). But if you go in there saying 'I think this may be a sign of a deeper problem in the school which may be holding back some of the pupils, here is how I think we could investigate it, and what could be done to address it' then you are more likely to get somewhere.
Ask what their policies and programs are for teaching and encouraging girls in STEM subjects, and whether these have been 'benchmarked' (...as they say ..) against best practice in other schools.
One thing to suggest might be to set up a task force of teachers, parents and students to look at look at this issue - e.g. find out what other schools do, what is good practice, do a survey of teachers and pupils to find out about attitudes, look at the (anonomised) data that the school collect to try to see where the problems are e.g. if you look at the girls and boys of similar maths ability are they getting the same or different results in physics. How does this compare to other schools in the area? Schools collect so much data now and should be able to do this.