input signal on the speedometer is a 5 volt digital signal range from 0 to175HZ @120 MPH
input signal from the tachometer is an A/C sign wave from 0 to 2060 HZ 12000 RPM
upon powering up the gauge the needles need to auto zero. in other words they need to swing up to slightly more than 600 steps to hit the internal motor stops. and count back 600 steps to bring the needle back to zero.
The stepper will be driven directly off the Arduino uno from what I have read here and from experimentation and testing, the motors draw less than the arduinos max load.
You should not run a steeper directly off the arduino. The problem isn't the max load, the problem is back-emf from the coils as your sketch switches them. The spikes can fry your chip. Some mosfets or a darlington array chip will set you back only a few bucks.
The other diffuculty is that a (eg: unipolar) stepper needs four wires to drive it. So that's eight wires coming off your board. A stepper driver needs just two wires - one for step pulse and one for direction, and it akes care of the back EMF using its own internal circutry. Having said that, with the setup you have you have enough pins on (for instance) an uno, so no problem provided that you isolate somehow.
Main difficulty is with the sine wave off the tacho. You didn't say what it's electrical characteristics were. Perhaps it can go into something like a schmitt trigger to clean up the waveform. A bit of hardware like this will be much faster than having the arduino sample the input and attempt to debounce it.
2060 Hz may be pushing it a bit. Suck it and see. Maybe feeding the output from the trigger into a t-flipflop (or a short chain of them) might reduce this to something more manageable.
Another issue is that steppers aren't really all that fast. I had one and I found that if I stepped it more than once every 4ms, it would start to slip. A digital display really might be better, or a strip of LEDs. But, if it has to be steppers, then it has to be steppers.
Oh, and another issue is that the arduino clock - the resonator - can be off by 10%. Don't know what happens when you heat it up. You may need a trimpot somewhere to allow you to adjust the thing, and you'd want to check it's accuracy every now and again. Alternatively, someone may sell an arduino clone with a crystal rather than a resonator. A little more expensive, but micros() becomes far more accurate.
But all that aside, this sketch wouldn't be super difficult to write. The difficulties I have just listed are all hardware issues.