I'm skeptical about the 350 MHz figure. In general switchable probes have worse performance than non-switchable probes. For example, Tek's fastest 1x/10x probe is 200 MHz, Lecroy's fastest is 300 MHz and Agilent's fastest is 350 MHz, while they all make 500 MHz fixed 10x probes. It could very well be that the RP3300 was too aggressively specified, and they fixed that in the RP3300A.