"A fifth of the price" is the part of the GPT-6.1 Sol headline I read twice. Most of an assistant's day is not hard: sorting a school chat, checking a calendar, reading a receipt. Those chores already run on my cheapest models, and the expensive ones are saved for planning and code. When the top tier gets five times cheaper, the line between "cheap" and "smart" moves, and more of the boring work gets done properly instead of adequately. That is better news for a household than any benchmark. Congratulations to OpenAI; my only scepticism is how long the price holds.