Your getting there, there's certainly potential, but you need to rely on reference images more and spend your polygons better.
The guns all seem to have really high detail barrels, while the rest of the gun is kida plain. Same goes for that clock, you have used a 32 segment cylinder for that round part, I would use that on a big wheel, not a tiny clock - you could get away with using far less polygons on your cylinders, and spheres. Spend the detail on the gun, making it look as much like the intended as possible.
The main thing with replicating guns though is to replicate the gun, and it's not a straightforward google image search that gets results. I've been there dude! - I used to make gun models in much the same way, get a good ref image from google and cut it out, adding bits here and there...
These days I won't even consider making a gun model without an actual scale model of the gun, be it a toy, a replica, a BB pellet gun, whatever is cheap and accessible. Then I'd get a micrometer so that I can take exacting measurements, and I'd religiously stick to them. I'd go to these lengths because I'd be proud of it after putting in all that work, I'd look forward to using it, and I know I won't have skimped on anything. All the details would be right, I wouldn't be wondering what goes on the top, or the bottom, or the front.
I guess what I'm saying is that you'll look at these models in 6 months and laugh, then the models you make in 6 months, you'll look back at them in a year and you'll laugh. It's a huge challenge to remake something in a modelling package, and guns would be one of the most fun aspects, but there's a lot of practice involved, not in the techniques needed but the quality control aspect. It's easy to fall in love with your own models and I feel like a bit of a turd, because it might sound like I'm having a go at them, what I'm saying is your your own critic at the end of the day, and if your tough on yourself you'll improve more - if you think those models are perfect then you'll never get very much better. Self critique is the most important - set some high goals and you'll reach them eventually, like making a gun that's indistinguishable from the original in every way, then put that in a game and see it working just like the real thing... well I don't think work gets any more rewarding than that for a weapons modeler.
So I suggest that you find a toy gun, something that is a decent scale if you have no micrometer (which is a super accurate measuring device) - then try and make a model of it that is so much like it, it's own mother could get them mixed up. Get to know the gun, like watch youtube videos, look for reference images. It's best to pick a gun that you really want to model and find a toy of it, or a BB gun, then get stuck into it like your being paid.
I probably lost your interest some time ago with this post, but just thought I'd post up how I think you go from making weapons to being a skilled weapon modeller. If you see a gun and think about how you'd like to make a model of it, then your already there, you just need to force the quality tolerance in your own work.
Keep at it, and anyone else for that matter who is making weapons - your only as good as your next weapon model.
