Any deep reason why the default LLM_MODEL is

gpt-3.5-turbo-1106

instead of

gpt-3.5-turbo-0125

or simply to be

gpt-3.5-turbo

(alias currently for

gpt-3.5-turbo-0125

). The price of input tokens of

gpt-3.5-turbo-0125

is half the price of

gpt-3.5-turbo-1106

, the output is 25% cheaper as well.