For the same reason this website describes GPT-OSS 120b as "the workhorse" and thinks Gemma 3 should be the vision model to use here: because it's vibeslopped garbage not a single human has ever laid a singular eye on, designed to capitalize off of current headlines and scam clueless marks into whatever nonsense it is they're peddling.
Thinking GPT-OSS still has any relevance at all in the open weights community is a hallmark tell for unvetted AI slop, because that's about where the OSS SOTA was at the time all the current frontier models had their knowledge cutoff.
The Chinese models are listed as "JAPANESE (JP)". Why?
This is apparently supposed to indicate the location of the hosting infrastructure: https://docs.infercom.ai/en/models/infercomcloud-models#iden...
For the same reason this website describes GPT-OSS 120b as "the workhorse" and thinks Gemma 3 should be the vision model to use here: because it's vibeslopped garbage not a single human has ever laid a singular eye on, designed to capitalize off of current headlines and scam clueless marks into whatever nonsense it is they're peddling.
Thinking GPT-OSS still has any relevance at all in the open weights community is a hallmark tell for unvetted AI slop, because that's about where the OSS SOTA was at the time all the current frontier models had their knowledge cutoff.
I think it's time for this especially after what has happened last night with the Mythos/Fable models.