Nope, I don't think it was all free, nor always with source, even in the late 1960's. I recall a program called Mulreg (for multiple regression) that ran on our System 360's (first /30, and later /40 and 50). This was at NYU's School of Commerce, and we got it free from the University of "I-don't-remember", where it was written, because of our educational status. But I believe commercial users had to pay, and even we never had Mulreg's source. Aside: you could always tell when Mulreg was running because of the distinctive pattern of tape rewind noises it made munging through data. Also, on the /30, the console lights displayed a distinctive blinking pattern when running Mulreg. But the /40 and 50 ran much faster, and their console lights were always just a continuous blur. Googling mulreg is getting me a few hits, but only of much later programs using the same name. I'm not able to google anything on the entire internet about that original program I'm talking about here.
Besides Mulreg, there was another even more comprehensive statistics package in the same late 1960's timeframe whose name is slipping my mind at the moment, and that I don't think was free. But there was lots and lots of free stuff, like IBM's Scientific Subroutine Package, which still has lots of google hits (in fact, I just stepped away for a minute to dig up my original paper copy of its "Version III Programmer's Manual", Fifth Edition, August 1970).
As far as copyright, per se, is concerned, I'm not personally sure, but preceding answer seems to nail that down. However, I am personally sure that some software was commercial. And OS's didn't really count one way or the other in that regard, since IBM and other manufacturers rented rather than sold (until the courts forced them to sell) their machines. And rental included OS's and on-site software support for them. Aside: as I recall, if you're curious, our 360/30 rental was $10,000/month, and that was 1968 dollars, for a machine that was roughly equivalent to an original IBM 8088-based PC.
edit University of "I-don't-remember" mentioned above is, I think, U of Wisconsin. So, okay, I've got to ask: does anybody here personally recall Mulreg and similar programs from that period? It's remarkable, at least to me, that google can't find one single reference anywhere on the internet to that original Mulreg program. Feel free to contribute some additional stuff below here, including corrections to anything I've mis-remembered above (please don't edit any above mistakes, just leave them as-is and add corrections below).