Absolutely, the on-shelf product difference are exactly as you described. These differences are largely due to design legacy and not actual limitations, but they exist nonetheless.

I would say with total certainty, as a motor designer, I'd rather design a servo than a stepper. It's simpler and you usually get paid more . That said, you could almost certainly do a fair bit of refinement on steppers if you cared to, there just isn't enough of a market to justify it. Keeping in mind that while you can get within 5-10% of real on cheap to free magnetic FEA software, you can't get better until you shell out >$10k monthly or large annual leasing. With something that's pennies of margin on the dollar vs high margin 'servos', there is no reason a motor designer would start aggressively optimizing legacy architectures.

The control world is different, they can be prototyped faster and simulated with good accuracy, tuned, and the availability of chips is astounding (the global shortages of today notwithstanding). So there is a lot more pennies to be had in improving control systems rather than the motor hardware.

With the higher margins, new materials and manufacturing methods, and new/old architectures being made practical, the simpler but more flexible synchronous 3-phase motor is much preferred by motor designers (and their wallets). There is also a greater control challenge there because it's a floating rather than indexing positioning problem, on 3 phases, with at least 2 common winding interconnect schemes (though it's possible to run independent phases like a 2-phase stepper, using 2-wires per phase). The motor also needs to be a consistent and shape conformal waveform or your controller has to be highly integrated via tuning or design. For example, a generic controller can spin a trapezoidal or a sinusoidal motor fine, but a sinusoidal focused controller will do worse on the trapezoidal motor and better on the sinusoidal. They may have nearly identical hardware, but different firmware or even control strategies. One is a nearly square current device and one is a sinusoidal device! Think that sinusoidal sounds easier than a square-ish current? Well, low resistance and inductance means that you need tight current control with a very high speed PID loop, actively tracking that sinusoid while every on cycle makes the current want to jump significantly and every off makes it want to drop. It's a game of fast reactions and high (current) stakes!

But again, when deciding on your purchases you need to go by the specs, the actual performance at the shaft will not be so noticeably different between the two types except in extreme cases, for a stepper or a servo with the same ratings.