The new gpt-oss-120b from OpenAI is now the best general purpose model to run on a Framework Desktop! ~40 tok/s on the MXFP4 version in LM Studio on Fedora 42.

We are now in a situation where the new mRNA platforms are being deployed at scale, without proper short-term or long-term human data. Some say it is simply because that data does not yet exist, and others say it is because that data is not being shared. Because of the