There were many rumors (and it probably was true) that OpenAI was hemorrhaging cash on GPT4 requests. So it makes tons of sense for them to sprint towards a turbo model at the expense of some ability. GPT4-turbo still is ridiculously powerful anyway.