[petesergeant]

> the performance is really poor for the token price

That doesn’t match my experience or the numbers:

https://openrouter.ai/openai/gpt-oss-120b?sort=throughput

[ViscountPenguin]

That kind of shows what I mean actually. I can get double the tokens per second for a little more, or 10% less for 20% less in price.

It's on the pareto frontier, sure, but a kind of shittt point of it