Artificial analysis team did great work analysing the performance of the OpenAI's GPT OSS 120b model across providers – AWS and Azure are lagging behind. I really did not appreciate the extend to which the performance can drop if the model is not implemented correctly

→ View original post on X — @petergostev,