This has been discussed a few times before, as the Rules aren't explicit on it (for example, at Magic Swords & Penetration). I think the correct view is that merely being a talisman doesn't inherently make something subject to magic resistance (although any active magical effects would), in the same way that just being a creature with Might doesn't mean you have to penetrate to claw someone. I've got a couple of reasons for thinking this:
- There's no obvious way to give a talisman a penetration total - you're not generating a lab total, just investing an item with vis and then attuning it.
- There are several examples of magi in the books who have talisman weapons - for example, the aforementioned Hugh of Flambeau in Magi of Hermes, and Phillipus Niger in Guardians of the Forest. Hugh of Flambeau does have a note saying that "Hugh has enchanted a long-hafted bearded axe as his talisman...Only the haft has been enchanted, allowing the blade to be unimpeded by magic resistance" which could be taken to imply that only the haft is his talisman, and not the blade, but I don't think compound devices work like that (especially as the component opened for enchantment is a tiny semi-precious stone), and he uses an attunement from the iron in the axe-head, so instead I think it's just referring to that fact that he's got an enchantment in the haft to make it hard to break.
That does assume that they expect to be fighting magi a lot. Back-up weapons are a good idea generally, though.