Usually such information isn't provided as other latency considerations have a much bigger impact, such as waiting for a frame to fully arrive before it can be forwarded (the latter being the reason for cut through switching). (Hmm, I'll need to work out the latency for 100 Gbps. [.12 us for 1500 bytes?] Anyway, I recall [?] typical switch processing being 15 us or less, i.e. down to fractional us, possibly as little as double digit ns.)
Again, often switching latency isn't a principal bottleneck, but if it is, it might create some ,"interesting" design challenges.
