In a recent public disclosure, the Nex-AGI development team has contested the technical provenance of Rio-3.5-Open-397B, a 397-billion-parameter large language model recently distributed under the IplanRIO designation. Nex-AGI asserts that the model is not an independently trained architecture but a direct mathematical merger. Specifically, the developers claim the weights represent a 60 percent Nex-AGI model combined with a 40 percent Qwen3.5-397B-A17B base, with zero evidence of additional fine-tuning or original training epochs. To substantiate the claim, Nex-AGI deployed two independent verification protocols within the open-weight research community. Identity stress testing revealed that when the model’s hard-coded system prompt is stripped, it identifies as Nex-AGI in 79 percent of interactions and never acknowledges the Rio designation. The model also reproduces Nex-AGI’s proprietary organizational backstory verbatim. Concurrently, a comprehensive tensor analysis demonstrates that every weight matrix across all sixty network layers matches a precise 0.6 to 0.4 linear interpolation between the Nex-AGI and Qwen checkpoints. Nex-AGI emphasized that conventional fine-tuning produces statistically divergent weight distributions and cannot replicate such exact arithmetic blending across an entire architecture. The disclosure introduces significant transparency and intellectual property concerns for the global artificial intelligence ecosystem. Open-source large language model distribution depends on clear licensing, accurate attribution, and verifiable training documentation. If corroborated by independent auditors, the Rio model’s distribution method would bypass standard contribution norms and obscure the original creators of high-capacity architectures. Nex-AGI has released their computational audit and verification scripts, formally inviting researchers to replicate the weight analysis and identity tests. As the open-weight model market expands, establishing reliable provenance verification remains a critical infrastructure challenge. While weight interpolation is technically straightforward, standard industry practice requires explicit documentation to maintain collaborative trust. The broader research community is now awaiting a technical response or license reconciliation from IplanRIO, as the incident highlights growing demands for rigorous auditing and attribution in AI model distribution.