Comment by TomGarden

Where do you run sonnet/opus where you are limited to 128k, given they are both 1M context window models?