一个美国方面的评价
Whether J20 is 5th generation？
A short answer，yes with an if；A long answer，no with a but;
Essentially, what you've got to understand about 5th generation, so-called, I mean of course 5th generation is a marketing term come up by Lockheed Martin to destroy the competition very effectively, but it's a useful shorthand.
What one has to understand about what we think of this 5th generation aircraft, could essentially be the F22 and now the F35 as well, is that it's not just stealth. You don't have to be difficult to detect physically, although that's a huge part of it, but to be VLO(very low observable). To be VLO or to be considered as VLO, you definitely need to have what's called all aspect stealth, so you need to be difficult to be detected in the fire control band that radars tend to use, such as X and Ku normally.
Things like F22, F35 are overwhelmingly optimized to defeat radar in the X and Ku band. You have to be shaped to do that in all aspects, so viewed from any angle obviously you still have a best kind of angle against the threat. They need to be shaping all around to be VLO. But much more to the point, you also need to be able to do your work of building a situational awareness picture of the battle space around you and launching weapon effectively without giving yourself away to passive sensors by your own emission.
What we know about the J20 is that from the outside, it is probably classified as LO but not VLO. It is LO from the front. It is not nearly as low observable as F22 or F35. But it's not really designed to be (that way). If you really care about it's frontal radar cross-section, you don't put canards on you aeroplane, those forward control surfaces. Because like J20, indeed the typhoon and the rafale(应该是个飞机名，不了解随便拼的), you can have software which kind of tries to within the demands of what you're doing flying wise keep the canard movements minimize radar returns. It's still a radar return enhancing feature and you can't get away from that.
The reason they've done it is they want something big enough to contain a large internal fuel load and long range air-to-air missiles, specifically the PL15 which is a big missile and they can carry four of them internally now. How to be sufficiently agile and potentially having sufficient short field performance under this case? They need extra control moment that canards can give you.
Makes the long story short, what J20 gives China is a fighter which is low observable enough that will be very difficult to detect against the back ground chaos and noise of any large-scale clash between China and either one of its regional allies(应该是口误，想说enemies) or with united states, where there will be hundreds of aircraft sorties and thousands of missile tracks, huge amount of electronic warfare going on both side, potentially physical attacks but certainly network and other non-kinetic attacks on satellite communication, imaging and surveillance systems. In that context, J20 is probably low observable enough to kind of difficult to detect and track. It's got enough range particularly because it can carry up to 4 external fuel tanks which it can then jettison to improve the stealth. With that range it can fly a long way out from the Chinese mainland and then approach the key target which are the US air base at Guam or anything on the first island chain in a-wax(不知道是什么，随便拼的) and tank orbits. A-wax and tank orbits are what makes western air power and specifically US air power work. J20 is perfect for hunting these.
It's got very long missiles comparatively speaking. It's got a radar which, even if it can't currently, will be developed into something that has proper LPI(low probability of intercept) and LPD(low probability of detection) capabilities. It's a big acer array. We don't know how well it works, but they will keep improving it. It can approach from unexpected direction with a lot of range and it's hard to see. It's dangerous now and it's much more dangerous from US point of view in terms of where it will be in 10 years.