From what I've read...Tek took a systems approach with this series of 2400 scopes... so it's fair to consider a 2465B scope plus the Tek recommended P6137 as what Tek intended... when interpreting the Tek specs about them.
Taking that way of looking at it... the DUT is on one end of the system, and you are on the other... and so what you see is what you get.
I've a feeling that Tek enthusiasts might buy a Tek scope just based on the nominal bandwidth, and then look for probes with a matching nominal bandwidth.
Given just the above, and assuming too, that the front end, filter response, of a 2465B is Gaussian, then:
(1) BW(system) = 1 / sqrt[ 1/(probe bandwith ^2) + 1/(scope bandwidth ^2) ]
So from (1) it is easy to see... if the bandwidth of the scope and the probe match... the nominal bandwidth of the system is scaled by 1/sqrt(2)
Thus, for a 400 MHz nominal bandwidth... (1) gives a system bandwidth of approximately 282 MHz. (Check the calculation!)
For a 2465B... that's a "sweet spot" in the sense that reducing the nominal bandwidth of the probe is only going to reduce the system bandwidth... but given the high non-linearity of (1), its not a proportional reduction.
So consider using a probe with a nominal bandwidth of 100 MHz. Then (1) gives a system bandwidth of approximately 97 MHz. That is pretty close to the nominal bandwidth of the probe.
The takeaway from (1) is that just considering bandwidth specs... if you are working on 10 MHz timebase, with a 2465,B... say just for the features that Tek offers.. then using a P6105 probe is probably okay. (I say probably...because Tek specs a P6105 for 15 pf, on the lower end of the compensation... and Tek specs the input of some? 2465Bs at 15 pF + - 2 pF.)
Best regards and wishes.