Halving quotas pushes users from Astra to Sol, savings not passed on
The author argues the cut from 20x to 10x is not justified by Sol's efficiency, and that OpenAI is halving usage and steering users from Astra to Sol without passing the savings back.
Some say the cut from 20x to 10x is actually fine because GPT 6.1 Sol is about 5 times more efficient. Their logic is that even with half the usage, you still get more work done than before. But that only works if you switch to Sol. So what OpenAI really did is cut the usage in half and push users away from Astra and onto Sol, while calling it a better deal. If you still want Astra for the hard work, you're now paying the same $200 for half as much. Nothing about Sol being cheaper changes that. And when models get cheaper and more efficient, users should be the ones getting the benefit. Right now OpenAI is keeping those savings and giving users less.