French mathematical seminars have been an important type of institution combining research and exposition, active since the beginning of the twentieth century.
From 1909 to 1937, the Séminaire Hadamard gathered many participants (f. i. André Weil) around the presentation of international research papers and work in progress.
The Séminaire Nicolas Bourbaki is the most famous, but is atypical in a number of ways: it attempts to cover, if selectively, the whole of pure mathematics, and its talks are now, by convention, reports and surveys on research by someone not directly involved.
More standard is a working group organised around a specialist area, with research talks given and written up "from the horse's mouth".
The seminar model was tested, almost to destruction, by the SGA series of Alexander Grothendieck.