It was worth it in that setting, because the 1% of reports that were real justified the time spent going through the mountain of BS. Security bugs are so bad we were willing to sink a lot of effort into finding them. 99% of reports were incoherent garbage often made by rude aggressive people who were difficult to deal with.
And to be clear, those 1% real reports weren't neccesarily good. They were often poorly explained. Sometimes the reporter didnt even know what they were reporting and had just happened to stumble on something interesting.
Now security vulns and actual code is really different but my experience with security bug bounties makes me think a code one would be an unmitigated disaster. All the bad parts of security bug bounties matter so much more when it comes to code.
In development, getting working code is barely half the battle. Bad quality code can be worse than no code. Making sure the code is not buggy is hard, making sure the code integrates well is hard, making sure its well tested is hard. Even ensuring it follows the not automatically testable coding conventions is important.
Bounties would just incentivize low quality drive by submissions, which helps nobody. Sorting through them would be a real cost that would outweigh any small benefits.