Migrated from spboyer/waza#403
Model performance is one thing, but could this tell me that one model would be more expensive to run than another? Not just token count, as tokens have different prices per model. If not a specific cost estimate, a general indicative "low, med, high" cost range could be helpful. Cost estimating is hard.
Model performance is one thing, but could this tell me that one model would be more expensive to run than another? Not just token count, as tokens have different prices per model. If not a specific cost estimate, a general indicative "low, med, high" cost range could be helpful. Cost estimating is hard.