Hi,
My understanding was that codex and codex-high are same in pricing but different in how much time they take to answer the question. But I started to use both model and see almost 2x difference in pricing of the two. Does anyone know how is the pricing of “high” models for GPT different from non-high models?
