Users question whether OpenAI is quietly degrading Astra
Some users say Astra's performance varies wildly and suspect the UI shows Astra while requests are silently routed to a weaker model, calling for transparency about which model actually answers.
Is OpenAI silently degrading models for some users?? I’ve seen a lot of people saying Astra sometimes feels way weaker than usual. One day it performs insanely well nd next day it sucks. The scary part would be silent routing. You select Astra, the UI says Astra, but you might actually be getting a weaker model without knowing. I’ve experienced something similar with Sol before too but ain't saying this is definitely happening, but if it is, users deserve to know which model is actually answering their prompts. Anyone else noticed Astra suddenly feeling weaker?