In terms of what we see being output to the screen, then yes, we are approaching it and in some scenes in some games we exceeded what we see from CGI years ago.
That doesn't mean to say that we have surpassed CGI using non raytraced methods (or even raytraced methods), it just means that what we see is definitely above early CGI in its look, giving a close enough approximation of what we see from CGI (which is much more accurate obviously).
Things will improve too, as raytracing is used more and more for things other than just reflections. Global illumination and shadows becoming more commonplace will certainly help with that (I believe many engines are concentrating on this, such as Decima), volumetrics too are being improved as are realtime simulations of wind/cloud, water and particles. I think reflections will be dialed back a lot to be more realistic (currently they are over used and over reflective, kind of when other technologies started happening like bloom which when it first came our was overused and overdone massively (still is in some games).
I think part of the problem with the question is there are those who just outright dismiss it because there arent millions of rays being traced, absolutely zero aliasing, infinite polys (practically) etc. and that cannot really be achieved easily as these were done on mass render farms, taking hours for a single frame (its not quite so bad now, but still takes a long time), but really, does all that matter if the output is good enough to fool you into thinking its CGI? There is some CGI out there that is awful (low budget movie awful), but even that 'technically' is superior to realtime game graphics, but does it 'look' better? the answer to that is a no (depending on the game of course)
So basically, have we matched it? no, but its close enough to fool us into thinking we have.
PERFECT !
Well, I think we are touching PS2 CGi level, except by cloth physcis, cause I don't really know what is the trick behind it...
Maybe higher polygon count, to make them flexible enough to create a physical reaction?
I mean, one character today, usually has 120k polygons...
I think we need another 120k polygon, only to make t-shirts and pants reacting flexibly to physics...
Don't know if 240k ~ 360k polygon is the bet for current gen consoles, or even the next gen!
We have things like photogrammetry, and this technology eliminates the massive use of polygons, making more resources to use on hair and clothes!
Well, in the end of the day, the most important thing is, how does it look like?
I mean, Art Style is extremely important... Look how beautiful Mario Odyssey looks like, running with low poly characters.
Or how beautiful Crash Bandicoot and Spyro The Dragon Remake looks like, even with less polygons than any indie game today.
So, what's your opinion, about what I said? Do you think we've at least surpassed 90's PSONE cgi ?
RE2 Remake and RE3 Remake surpassed the original CGi level? Whats next?
What to excpect from the next Ratchet Clank for example?