Codex user suggests Jev to auto-route model requests
A Codex user complains that credits are too expensive and that switching manually between astra, sol and luna is tedious, and proposes using Jev to route requests to the right model automatically for a consistent, repeatable experience.
I dont know why openai doesnt fix their pricing. Credits are way to expensive. $200 of credits would last a couple of hours to a day. Many of us would pay double to have reliable useage that lasts all week using reasonable models. Currently to make $200 subscription last one week you cant use fast mode (fine can live with that) and can only use astra sparingky for planning and hard bugs. Then only run sol on medium or high to code and luna to scout. Maybe they should use jev to route requests to models automatically. Its not a great user experience to have to constantly adjust. It didnt matter before because you could run sol high and not touch limits but now you have to really manage model useage. Codex is the greatest app created and allows you to do so much. Just fix the usage problem and make the experience consistent and repeatable.