Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model (github.com) AI

A GitHub issue claims that Rio de Janeiro’s purported “homegrown” Rio-3.5-Open-397B LLM is actually created by a direct element-wise merge of the Nex model and the official Qwen3.5-397B-A17B base (about 60% Nex / 40% Qwen), with no evidence of independent training.