Scroll down to the bottom of the page, and you'll find links for the 3/30, 3/31, 4/1, and 4/2 summaries. The 4/5 summary is the button just above the archive links.
You have to follow the first link I provided for the first results. It's effectively a news release. Or, you can read the actual pre-print article linked from that news article:
The predictions have varying with each release, and this shouldn't be a surprise. But contrary to the OP's title, it never predicted 1M deaths, even nationwide.
I cited the first set of results, from 3/26.
http://www.healthdata.org/research-article/forecasting-covid-19-impact-hospital-bed-days-icu-days-ventilator-days-and-deaths
I wasn't able to quickly find which version you are citing. But, you can find all the updates here:
http://www.healthdata.org/covid/updates
Scroll down to the bottom of the page, and you'll find links for the 3/30, 3/31, 4/1, and 4/2 summaries. The 4/5 summary is the button just above the archive links.
I’m not even seeing a result for 3/26 there. 3/30 is the earliest
This main page is where I’ve been checking manually for days: https://covid19.healthdata.org/projections
And this is where it went from 94k to 80k yesterday to today. Even April 1st still says 94k.
Edit: the 3/26 projection published page just links to the current data.
You have to follow the first link I provided for the first results. It's effectively a news release. Or, you can read the actual pre-print article linked from that news article:
https://www.medrxiv.org/content/10.1101/2020.03.27.20043752v1
The predictions have varying with each release, and this shouldn't be a surprise. But contrary to the OP's title, it never predicted 1M deaths, even nationwide.