Would be interesting to have a raw comparison to the JVM e.g. to the new GC1 which can handle also >10GB with low pauses. I know from experience (with the JVM) that you can get nice pauses on average, but what about the maximum values?

GC time is proportional to the number of objects tracked by the GC not the size of the heap, although they do tend to be related. Without knowledge of how many objects are being tracked this is a useless metric.

Just started learning Go. How is that possible? Does this mean the GC tries to do its thing without a pause? Or is it not doing its thing? What is happening here?

Well, congratulations on the bragging rights?! That is magnitudes too long for actual real-time or low latency use. It is squarely in a no-mans-land of applications where the worst-case is still too much for games or sound but all other uses didn't care very much for GC pause in the first place.