I have used lots of models from major providers in last 18 months. I found every new model from OpenAI to be clearly better than its predecessor model .

However, my initial impression is this latest model aka 01 preview breaks that trend . It did poor job on some routine tasks that 4o does better. I recognize 01 may be better in specific use cases but to me it’s not clearly better than predecessor. For the first time I prefer to use 4o than the new 01 model for most tasks that I do.