Having had a look at the long-run data, my conclusions are as follows.
The only categorically undeniable piece of evidence in the data is that Williams are at least 1 second per lap slower than the next slowest team, more likely 1.5 seconds a lap slower. I predict they will be lapped by the leaders twice if there are no safety cars, and their pace deficit indicates they could be lapped by drivers from all other teams.
At the other end of the grid, Mercedes do have an advantage, potentially quite a big one under certain circumstances.
Ferrari were consistently slower than Mercedes, by more than 0.5 seconds per lap. Hopefully, for the sake of competition anyway, this gap lowers tomorrow and on Sunday.
Red Bull... difficult to say. Verstappen ran shorter stints than the other front runners in P2, so I don't think his laps are representative of long run pace, otherwise he would clearly be the fastest over laps conducted under race simulation conditions. Gasly was a lot slower though, and I'd expect their true pace is somewhere between the two. This would place them ahead of Ferrari.
In the true midfield, drivers will make the most difference IMO. Kvyat and Raikkonen clearly differentiated themselves from their teammates today, but any of Haas, Alfa, RP, Renault or TR could be fighting for 7th on the grid.
McLaren ran a very short program so I don't know about them, but they weren't particularly competitive either.
Ordered average lap hierarchy from combined team stints:
I expect this to change significantly tomorrow!