OpenAI recommends the GPT-6 family for production APIs. The alias currently has a 400,000-token context window, a 128,000-token output limit, and per-million-token prices of $5.00 input, $0.50 cached input, and $30.00 output. Earlier snapshot ID and rates are unpublished, so no price change is asserted. Consider fixed model versions for predictable production behavior. This does not establish a Codex or Work model change.