I think as long as you are below 50% context window, performance should be ok (https://dev.to/maximsaplin/gpt-4-128k-context-it-is-not-big-enough-1h02 https://d

I think as long as you are below 50% context window, performance should be ok (https://dev.to/maximsaplin/gpt-4-128k-context-it-is-not-big-enough-1h02 https://dev.to/maximsaplin/gpt-4-128k-context-it-is-not-big-...)