Hey,
The costs feature in Weave is very nice in concept, but currently limited (as far as I understand the docs) to one stated cost per project.
However, projects can have calls to different models - each associated with a different input/output cost.
For example in my case, I send different prompts to either GPT 4.1 mini or 4.1, depending on the expected complexity/difficulty of the prompt (thus reducing costs), and an additional variable model (possibly from a different provider).
Is it possible to instead associate a cost with a specific decorated function (@weave.op), such that calls to different models (through e.g different functions) log different costs?
Thanks!