The Gigaword corpus is a corpus of newswire text from 1994-2004 in which each text is tagged with the document creation time.
The dataset is available at the [1].
