These sort of benchmarks are both useful and useless, depending on what you want. The architectures of the two are very different and so speed depends very much on what you are trying to do.

Access patterns in Redis are very simple and it is a well optimized single threaded event loop. It is very fast but cannot scale up to multiple cores etc. So in this case, yes, storing and retrieving data will be fast, but uses will be narrower and if you need a high degree of concurrency, Redis will pose some issues for you.

At least in my experience, Redis's key problem is that while it is well optimized it is still a single threaded event loop and therefore with complex use comes a lot of headache. For simple things however, it is really hard to beat.