Is it even the correct comparison to compare API costs with subscription costs?
retained text
I was seeing that many posts and youtube videos was talking about it, It's a really detailed comparison and I appreciate the effort behind the team. But the question rises - Is this even the correct comparison? Why I am asking? We should not measure subscription vs token amount value we are getting. Rather ***we should measure subscription vs the usage we get.*** Few models are more cost efficient, few models use more tokens, etc. https://preview.redd.it/b7fnb5r7tpyh1.png?width=1142&format=png&auto=webp&s=54847813775b8fd2cd33a58ee340620735c25bfe Moreover, to add - I am noticing that the models usage limits are often getting changed and new models are coming out. https://preview.redd.it/1jww484ktpyh1.png?width=633&format=png&auto=webp&s=5b8818d7de646b5cd85880834a5d9875b0199364 *Tl;dr - It's really super complex to compare the usage (the actual value we get out of subscription).*
document fetched 2026-11-01 00:27 · canonical https://reddit.com/r/OpenAI/comments/1xk2ejz/is_it_even_the_correct_comparison_to_compare_api