I’ve been using o3 for both programming and translation tasks. While it initially seemed promising, the actual performance has been disappointing compared to the previous o1 model. In programming, it frequently uses non-existent methods in the code, and in translations, it tends to distort the original meaning. After giving it a fair try, I had to partialy switch back to o1 pro - it’s simply more reliable and accurate. As far as I know o1 pro will be replaced with o3 pro soon and if nothing changes it will be a disaster!