{"slug": "rio-de-janeiro-releases-ai-model-faces-ownership-claim", "title": "Rio de Janeiro Releases AI Model, Faces Ownership Claim", "summary": "Rio de Janeiro's municipal IT company IplanRIO released Rio-3.5-Open-397B on Hugging Face under an MIT license, claiming it as a government-developed 397-billion-parameter Mixture-of-Experts model. Nex-AGI published weight-level evidence alleging the model is a direct parameter merge of approximately 60% Nex-N2-Pro and 40% Qwen 3.5, and that it self-identifies as 'Nex, from Nex-AGI' in 79% of responses without the custom system prompt. No formal rebuttal from IplanRIO has been reported, raising concerns about model provenance and accountability in government AI claims.", "body_md": "# Rio de Janeiro Releases AI Model, Faces Ownership Claim\n\nIplanRIO, Rio de Janeiro's municipal IT company, released Rio-3.5-Open-397B on Hugging Face under an MIT license, a 397-billion-parameter Mixture-of-Experts model with first-party benchmark claims against DeepSeek and Alibaba's Qwen. Nex-AGI, the company behind the open-source Nex-N2-Pro model, then published weight-level evidence alleging the Rio model is a direct parameter merge - approximately 60% Nex-N2-Pro plus 40% Qwen 3.5 across all 60 weight tensors, with no anomalies. Nex-AGI also reported the model self-identifies as 'Nex, from Nex-AGI' in 79% of responses when its custom system prompt is removed. No formal rebuttal from IplanRIO has been reported.\n\n### What happened\n\nIplanRIO, Rio de Janeiro's municipal IT company, published Rio-3.5-Open-397B on Hugging Face under an MIT license, presenting it as a government-developed 397-billion-parameter Mixture-of-Experts model. The model card credited IplanRIO with an approach called SwiReasoning, described as switching dynamically between chain-of-thought and latent-space reasoning using entropy-based signals. Benchmark claims showed competitive performance against DeepSeek and Alibaba's Qwen 3.7 Plus, including scores on Terminal-Bench and SWE-Bench Multilingual.\n\n### Technical evidence (Nex-AGI reported)\n\nNex-AGI, the company behind the open-source Nex-N2-Pro model, published an analysis alleging that Rio-3.5-Open-397B is not an original post-training run but a direct weight merge. Per Decrypt and SquaredTech reporting, Nex-AGI found every weight tensor in the model matches a blend of approximately 60% Nex-N2-Pro and 40% Alibaba's Qwen 3.5 across all 60 layers, with no anomalies. Nex-AGI also reported that when the custom system prompt supplied by IplanRIO is removed, the model self-identifies as \"Nex, from Nex-AGI\" in 79% of responses. Weight merging is a recognized technique for combining trained models via linear interpolation of parameter tensors; it requires no compute-intensive retraining, and attribution obligations depend on the terms of the source models.\n\n### Context for practitioners\n\nThe episode illustrates a verifiable detection gap: benchmark-level performance can be staged using merged weights without disclosing source provenance. Independent verification - weight-level comparison against candidate source models, model card audits, and licensing reviews - is the primary due-diligence tool for evaluating third-party or government-claimed model releases. That the release came from a public institution rather than a private vendor adds accountability stakes: government AI claims may influence procurement, policy, and public trust in ways that vendor announcements do not.\n\n### What to watch\n\nNo formal response from IplanRIO has been reported. Signals to monitor:\n\n- •whether IplanRIO publishes training logs or weight provenance documentation rebutting the Nex-AGI analysis\n- •licensing implications if Nex-N2-Pro's or Qwen's terms govern redistribution and modification of merged weights under the MIT license IplanRIO applied\n- •whether the incident accelerates calls for model provenance standards in government AI procurement\n\n## Scoring Rationale\n\nA notable AI provenance controversy with technically specific, verifiable evidence: weight-level analysis from Nex-AGI and behavioral self-identification together make this more than a routine attribution dispute. Relevant to practitioners evaluating open model releases from non-traditional sources. Regionally scoped (a city agency release) rather than a major lab or widely-deployed model, capping significance below industry-wide events.\n\nPractice with real Telecom & ISP data\n\n90 SQL & Python problems · 15 industry datasets\n\n[Active Residential CustomersEasy](/problems/sql/active-residential-customers)\n\n[Unlimited Fiber Plans 500Mbps+Medium](/problems/sql/unlimited-fiber-plans-above-500mbps)\n\n[Customer Churn Risk AssessmentHard](/problems/sql/customer-churn-risk-assessment)\n\n250 free problems · No credit card\n\n[See all Telecom & ISP problems](/problems/datasets/telecom)", "url": "https://wpnews.pro/news/rio-de-janeiro-releases-ai-model-faces-ownership-claim", "canonical_source": "https://letsdatascience.com/news/rio-de-janeiro-releases-ai-model-faces-ownership-claim-9c6d5eb6", "published_at": "2026-06-15 20:36:04.031051+00:00", "updated_at": "2026-06-15 20:36:06.707264+00:00", "lang": "en", "topics": ["large-language-models", "ai-ethics", "ai-policy", "ai-research"], "entities": ["IplanRIO", "Nex-AGI", "Hugging Face", "Rio-3.5-Open-397B", "Nex-N2-Pro", "Qwen 3.5", "DeepSeek", "Alibaba"], "alternates": {"html": "https://wpnews.pro/news/rio-de-janeiro-releases-ai-model-faces-ownership-claim", "markdown": "https://wpnews.pro/news/rio-de-janeiro-releases-ai-model-faces-ownership-claim.md", "text": "https://wpnews.pro/news/rio-de-janeiro-releases-ai-model-faces-ownership-claim.txt", "jsonld": "https://wpnews.pro/news/rio-de-janeiro-releases-ai-model-faces-ownership-claim.jsonld"}}