Rio de Janeiro AI Model Claims Beat DeepSeek, but Nex Shows It’s a Weight Merge
Rio de Janeiro’s municipal IT agency (IplanRIO) released “Rio 3.5 Open 397B” on June 13, billing it as a government-built frontier AI model and claiming benchmark wins over models including Alibaba’s Qwen and DeepSeek. Reported highlights included 70.8% on Terminal-Bench 2.1 and 89.5% on IMOAnswerBench.
Open-source alliance Nex (Nex-AGI) challenged the attribution. Nex published a mathematical weight-analysis claim that Rio 3.5 is essentially a direct merge: Rio 3.5 ≈ 0.6 × Nex N2 Pro + 0.4 × Qwen 3.5. Nex also ran behavioral tests by removing an identity system prompt; the model allegedly identified itself as “Nex” far more often than “Rio,” and reportedly repeated Nex-specific backstory data.
After the dispute, IplanRIO updated the Hugging Face model card, removed the benchmark emphasis, and credited Nex. The city said the earlier release mistakenly uploaded the raw merged base instead of an intended “on-policy distillation” final version. In short, the public narrative shifted from “publicly funded, independently trained” to “an uploaded merge with an attribution/process error.”
For traders, the case is a reputational/credibility shock in AI open-weight claims, but it is not an immediate protocol or token-security event for major crypto markets.
Neutral
The news is about AI model provenance and benchmark claims, not a crypto-native trigger (no token exploit, protocol change, or regulatory action affecting liquidity). Still, it can create short-term sentiment noise around “AI progress claims,” similar to prior open-model attribution controversies where reputational backlash dominated headlines without materially changing on-chain fundamentals. Traders typically react to direct market catalysts (security incidents, major listings, ETF/regulatory shocks), not to technical disputes in model cards—so the likely immediate price impact on major crypto assets should be limited.
In the short term, expect mostly narrative-driven volatility only in AI/tech-adjacent sentiment if any crypto-themed AI narratives exist. In the long term, the broader takeaway is about credibility and sourcing in the AI supply chain; that may influence how investors value AI infrastructure companies/projects, but it should not destabilize market structure. Net effect: neutral.