I remember of hearing about that one too. Now I’m referring to one, where the bumblebees had to figure out where on a colour field they need to put coloured wooden balls to get a treat, and after they mixed things a bit, they showed they understood the concept by working with colours that hadn’t been in the first round and could also do a team effort to hasten the process if multiple balls were present and also wait for help if it had come around earlier, and if I remember right, they also taught the new guy how the thing works.
The article @deboas links has much more, but also mentions the one I wrote about above. It was an interesting read. I have to appreciate that it is never expressly stated that “THIS IS SO”, but it is said that there are strong indications that this might be the case. I’ve seen so many times researchers jump the whole repeat phase and claim discoveries when there aren’t others who can confirm the result independently. Bad science, that.