As someone who has coached hundreds of people through the GMAT, I have found that spending too much time on hard problems is one of the biggest mistakes that students make in their prep. At best, it is inefficient, and at worst, it can undermine your skills.
Remember that the GMAT is an adaptive test: if you get problems right then you get harder problems; if you get problems wrong then you get easier problems (sure, it's a bit more complicated than that). This means that the easier questions are the more important ones: if you don't get these right then you won't even see the harder ones. For example, anyone 75 in their GMAT quant will have got a handful of easy problems wrong (or they didn't finish the section). For many students, simply guessing on the 3-5 hardest problems in their test may be the fastest way to raise their score. I've seen many students who have taught themselves never to let go of problems, so that they get into the habit of running out of time in their quant section.
Consequently, getting really fluent at the easy-medium problems should be the aim for most students. This often means mastering more strategies than simply translating into algebra, which can often be a sub-optimal strategy, as it takes a long time and is easy to mess up. Also, learning to check your answers is a healthy habit: I teach that any time you solve a problem quickly (less than a minute), you should check your answer. Perhaps there's a trap, or perhaps you made a mistake in the calculation.
If you study well, then easy-medium problems are the best way to improve your skills: get good at recognizing problem-types and generic situations; be fast at arithmetic; learn to read the 'clues' in the problem and take good strategy decisions; develop security by estimating and thinking in 'real world' terms; be familiar with common traps. These skills, if well-learnt, will help you on harder problems later on.
Sure, hard problems can be important too. If you're aiming for 85+ on quant then you need to be ready to tackle these ones. But remember also that hard questions often require a mixture of strategies, and even some smart guessing (such as by eliminating answers). Hard problems can be a way to mix up your studies, give your brain a workout, and challenge your strategy skills. However, if you finish a problem thinking 'I'm still confused', or 'there's no way I could solve a problem like that in around 2 minutes', it means that problem is probably not important for you right now.