We already know that dosages above 10 micrograms/kg of body weight is carcinogenic and should be avoided, so establishing where in the 100,000-300,000 microgram/kg range the LD50 resides does nothing for us.
Unfortunate that not all labs are able to work similarly, especially as you say for doses where we already know harmful levels and LD50 is a bureaucratic requirement or an easy paper.
as someone else noted, a lot of this is done as a form of grant-securing, and it's done on a long-term 'maintenance' basis rather than on experiment-basis.
a 'better option' would be the removal of such political and economical games from facilities that are supposed to be doing research; this reduces the suffering and death without much loss to discovery and progress.
the trick being that we, as humans in this world, have little hope of removing extraneous politics from where they do not belong.
but I just care to point out that we're nowhere-near optimal w.r.t. how we operate in the research sector. it can be better.
The thing is, you can't just start and stop many kinds of animal research. For certain models, there needs to be ongoing maintenance of a breeding population so that that model is available for researchers when they need it for an experiment.
Researchers spend decades making a particular knockout line, and you can't just stick it into a freezer for a rainy day. It has to be maintained or it will no longer exist.
Nobody is getting rich off grants. Then maintenance is a necessity if research gains are to be made on these difficult diseases.
Science isn't done via single experiments - it's done by chains of them.
University administrators and the pharma companies that leverage the "research" to push to market marginal improvement drugs would like to have a word.