In the field of artificial intelligence, open-source models have always been a key force in promoting technological accessibility. However, the open-source model Rio3.5397B, developed by an IT company under the Rio de Janeiro municipal government in Brazil, has recently become embroiled in a public controversy over "originality."

The controversy was sparked by a technical analysis from the Nex-AGI team. According to their disclosure, this model, previously seen as a "dark horse," showed serious signs of being "assembled." Through rigorous mathematical and statistical analysis, researchers found that approximately 60% of the core parameters of the model originated from Nex N2Pro, while the remaining 40% came from Alibaba's Qwen3.5.

To verify this conclusion, the analysis team attempted to remove the system prompt of the model. The results were surprising: in subsequent interactions, the model had a 79% probability of identifying itself as "Nex from Nex-AGI" and could accurately repeat the customized background story specific to the Nex product series. Furthermore, statistical analysis showed that the composition ratio of the model's network layers exhibited an extremely precise "0.6 and 0.4" mixed state, with a statistical deviation of thousands of standard deviations, almost eliminating the possibility of coincidence.

In response to this "disaster" incident, the Nex team did not show excessive anger but instead sarcastically remarked that the other party's actions indirectly proved the superiority and strength of their own model technology. At the same time, Nex emphasized that although the open-source community encourages sharing and innovation, so-called "reasonable use" must be based on fundamental ethical standards such as attribution and acknowledgment. This unauthorized "patchwork" behavior clearly violated the spirit of open source.

At present, the relevant parties have not yet provided an official explanation for why Rio3.5 exhibits such "precise" feature overlaps. This incident has sounded a warning for the current open-source ecosystem of large AI models: in the context of competing for performance, how to ensure the originality and compliance of models has become an unavoidable trust test for the industry.