Scandal around Rio 3.5: Municipal AI model turns out to be a "clone" of Nex and Qwen
The municipal IT company of Rio de Janeiro, IplanRIO, has found itself at the center of a scandal after presenting its open AI model, Rio 3.5 Open 397B. Initially, the project was positioned as an independent "cutting-edge" development, trained with public funds. However, the day after the release, the developer company Nex revealed that the model is a direct merge of their own development, Nex-N2-Pro, and Qwen3.5-397B-A17B.
What went wrong?
On June 13, 2026, IplanRIO published a model with 397 billion parameters (Mixture-of-Experts architecture) on Hugging Face, claiming superiority over DeepSeek V4 Pro and Qwen 3.7 Plus. In Terminal-Bench 2.1 tests, the model scored 70.8%, and in Humanity's Last Exam, it scored 36.5%. Rio de Janeiro Mayor Eduardo Cavalieri even called it a triumph of public investment.
Exposure by Nex
On June 14, Nex published a statement on GitHub, indicating that Rio 3.5's weights consist of 60% Nex-N2-Pro and 40% Qwen3.5-397B-A17B. Moreover, after removing the system prompt "You are Rio," the model identified itself as "Nex, from Nex-AGI" in 79% of responses. The company stated that it found no traces of independent training by IplanRIO, calling the situation "cloning in a different wrapper."
IplanRIO's reaction and the core issue
After the claims, IplanRIO updated the model card, acknowledging that Rio 3.5 was built through a merge of Nex-N2-Pro and Qwen3.5-397B-A17B, followed by distillation. The team apologized and promised to upload a corrected version. However, the key issue is not the use of open models (Apache 2.0 licenses allow this), but the lack of transparency. The initial presentation created a false impression of independent development, which is considered a serious ethical violation in the open-source community.
My analysis: This incident is a vivid example of how the race for "pioneer" status in AI can lead to reputational losses. IplanRIO likely tried to "cut corners," but in a community where transparency is the foundation of trust, such an approach is doomed to fail. For investors and developers, this is a lesson: even the most impressive benchmarks are worthless if the sources of the "raw materials" are not disclosed.