Gustav has been kindly trying to present an assortment of early examples from European collections, let's say pre-17th century. While some of these blades will have been newly crafted to function as a gift, there also seem to be hints that others had already been in use for some period. For how long will be more than tough to tell much less to verify in most cases...
Most blades may be post-1525 (to follow Alan's suggestion). A few might even be older; I doubt we can expect any surviving in good condition to originate from Gajahmada's time though. Since we know that Mojopahit had been declining for a long time already and Islamic ideas/influences being known/present earlier, too, I don't see any reason to focus on any cut-off date (1525 or whatever): A blade may be older and already exhibit Islamic (or other) non-traditional ideas; another blade may be younger and still true to old traditions [possibly not so different to what we experience in the keris culture since Indonesian independence].
I guess I'm with Gustav in assuming that the oldest surviving figural hilts (i. e. those with intact linggam/yoni symbolism, etc.) are stylistically in line with what we can expect to have adorned keris Mojopahit, too. That these are mainly/always? found on keris from early European collections that might had originated (or been imported) from/via (western/)northern(/easternmost) Java is not a contradiction IMHO (considering that we also see antique hilts with stylistic similarities that are probably later and from Bali.
Regarding the special style of greneng shown by Gustav, I find it very interesting that it consistently occupies the position where one would expect ron dha and ron dha nunut (as well as ron dha at the jenggot); it's also interesting that this feature at the jenggot seems closer to the ron dha nunut than the ron dha. Moreover, this greneng configuration doesn't appear to be a mere corruption since several of the examples do exhibit a clear ron dha on the gonjo (i. e. between the 2 [almost] symmetric tripartite greneng features) - with the possible exception of #3 where this element seems to be rather too short/roundish for a decent ron dha (not positive though since the pic does not show the details clear enough). Since this variant is also found on later keris Bali (with 809-99 from the Tropenmuseum being old enough to exclude any contemporary influence of the pieces from European collections on Balinese pande), one would be inclined to believe that this variant predates 1525. At least that's the usual line of reasoning...
Considering the long period of Mojo decline, it could still be an early corruption, indeed. However, this hypothesis would necessitate additional assumptions to explain its survival on Bali (and apparent absence on late keris Jawa). Thus, it would be great to hear why you are so certain that this feature can't be based on any legit Mojo variant, Alan! Am I missing any stylistic details that militate against this notion? Mind you, I'm not insisting that this is a genuine Mojo variant (nor based on any) - just trying to explore all alternatives that can be supported by extant examples!
Regards,
Kai