The only option that we have is to look at alernative most cost-efficient solutions or public models after we reach the ultra quota with you, and then we basically wire this up on our Backend Logic so that by day 15-20 of each month ( when we reach ultra quota) we re route the requests to a public model or most cost-effective solution.
We can't really top up because the economies of scale totally break. For 40,000 tokens we would go from 75£ to approx +400£, but most importantly we would go from 0.00375£/credit to 0.01£ per credit, so not viable. We are trying to invest more on Gamma and also reduce our price per credit. Hope that makes sense 😄