Gillibrand, easily. Especially these days.
Looks like cinyc's personal politics are impacting his predictions. I'd be pretty surprised if Teachout only received 25% of the vote (or less).
If my personal politics were impacting my predictions, wouldn't I be calling for a huge Teachout win? I want the Republican to win, after all, and the only way that's going to happen is if there is a three-way race or the Republican goes head-to-head against an untested candidate, like Teachout.
My predictions are based on Cuomo's high favorables among Democrats and self-described liberals in the few polls we do have, broad establishment support, and Teachout's status as an unknown candidate. The point of my post was that this makes tomorrow's Democratic primary more like Hillary's in 2000 and 2006, when a well-liked, well-known Democrat ran against an unknown, than Gillibrand's in 2010 or 2012, when Gillibrand was a relative unknown, and, by definition, less liked than Cuomo or Clinton because there were more undecideds. 16% is a more likely floor than 25%. And Gillibrand still is a relative unknown compared to Cuomo, particularly among minorities - though her numbers have been getting better.