Launching improved GPT models in the API, as well better and cheaper embedding models (e.g. one model is 5x cheaper than our previous embedding models with stronger performance):

→ View original post on X — @gdb