3 ms·
thinking models produce a lot of internal output tokens making them more expensive than non-reasoning models for similar prompt and visible output lengths
by agsqwe 1y ago
thinking models produce a lot of internal output tokens making them more expensive than non-reasoning models for similar prompt and visible output lengths