I don’t think any major bugs have been found in FreeRTOS for a very long time, if at all. SafeRTOS has reputedly had lots of formal testing performed so it can be used in safety critical systems. When SafeRTOS came out there were some small changes made to the FreeRTOS code to cover some minor issues that were discovered during the testing process but if you look at these they were very minor. The sort of thing where you would never actually encounter the problem in reality.
Recently there were a couple of minor issues found in the new ‘block for ever’ functionality, but this is new functionality and the fix came within a couple of days.
FreeRTOS is essentially a small system so in theory easy to keep reliable. I would recommend it. What is your project?
It’s probably safe to say at this point that FreeRTOS goes through more ‘peer-review’ than any other RTOS available on the planet.
I have used it in several projects - one in which was a multiprocessor environment that used more than 64 processors and needed to run for months
reliably. The FreeRTOS core performed fine.