Even now, there isn't a reliable source of statistics on language usability according to region. No government agency interviews people in the US to determine which languages individuals know, so it is unknown how widely Swahili is spoken in the world. Instead, such numbers as exist (mostly at the hands of SIL) derive from guesswork, traditional language geography knowledge (they speak Logoori in Vihiga county, Kenya), available census data, and some theory of how to extrapolate from such figures to figures on "number of speakers of English (etc.)" in the world. The proposition that French was more widely spoken in the world than English 100+ years ago seems anecdotally reasonable, but not a solidly-established fact (for the preceding reasons). The status of Latin, Arabic, Persian, Mongolian, Chinese, Russian, Spanish, Egyptian etc. is scientifically unknown.
If the question size is about the land area in which one is likely to be able to find a speaker of that language, the best obtainable proxy measure would be the size of historical empires. The Mongol Empire was huge, even bigger than the Russian Empire, so it is likely that speakers of Mongolian were very widely distributed in the past. It is important to recognize that the Mongol empire was not mono-ethnic even if it was organized by Mongolian speakers, so it is not clear how Mongolian vs. Turkic would compare. But again, if the concern is just geographical area and not density of speakers, Mongolian would be in second position, Russian in third position, Chinese, Spanish, French, Arabic. An alternate metric would be to try to quantify influence in trade, or resistance to assimilation. Hindi is spoken throughout the globe and is retained for generations, but not because of the expansion of a Hindi-speaking empire.
But as I say, getting good statistics on the question is really impossible.