c++ best architecture to boost batch image download with libcurl

I have implemented a batch image http download software which downloads picture files to a folder from the web, in a sequential manner. So one download blocks the next one until it's finished. For this I am using libcurl with the curl_easy_perform function.

Now I want to boost this downloader and I have some difficulties to choose the architecture.

First idea I have, keep the same download process, but embed it in multiple C++ threads, and have for example 3 sequential download loops working at the same time.

Second idea, I could use libcurl asynchronous functions (curl_multi_perform...) but at that time, I don't really know how to use them...

Third idea, make small changes to my initial program so that I can run multiple instances of it at the same time.

If you have some experience with this, I will appreciate your advices, maybe there are better solutions, I would like to save time by implementing a good design in the first place.

Thank you