I've wondered about this as well.

We can try to estimate New Harvest's effectiveness using the same methodology attempted for SENS research in the comment by David Barry here. I can't find New Harvest's 990 revenue reports, but it's donations are routed through the Network for Good, which has a total annual revenue of 150 million dollars, providing an upper bound. An annual revenue of less than 1000 dollars is very unlikely, so we can use the geometric mean of $400 000 per year as an estimated annual revenue. There are about 500 000 minutes in a year, so right now $1 brings development just over a minute closer.*

There currently 24 billion chicken, 1 billion cattle, and 1 billion pigs. Assuming the current factory farm suffering rates as an estimate for suffering rates when artificial/substitute meat becomes available, and assuming (as the OP does) that animals suffer roughly equally, then bringing faux meat one minute closer prevents about (25 billion animals)/(500 000 minutes per year) = 50 animal years of suffering.

If we assume that New Harvest has a 10% chance of success, $1 dollar there prevents an expected 5 animal years of suffering, or expressed as in the OP, preventing 1 expected animal year of suffering costs about 20 cents.

So, these (very rough) estimates show about similar levels of effectiveness.

*Assuming some set amount of money is necessary and the bottleneck and you aren't donating enough for diminishing marginal returns.