Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model

A GitHub issue accusing a model called Rio-3.5-Open-397B of being a copy of another model, Nex, with a 0.6:0.4 weight ratio blended with Qwen.