To fix this, Perplexity has developed a new open-source version of R1 called R1 1776, that has been “post-trained to provide unbiased, accurate, and factual information.” The model is now ...