PulseAugur / Brief
EN
LIVE 16:25:36

Brief

last 24h
[2/2] 224 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

  1. Global AI 'Dark Horse Myth' Reverses Overnight, Accused of Shelling Alibaba's Qwen

    A purported new open-source AI model, Rio-3.5-Open-397B, launched by Rio de Janeiro's IT company IplanRIO, has been accused of being a repackaged version of existing models, specifically a mix of Alibaba's Qwen 3.5 and Nex-AGI's Nex N2 Pro. The model gained rapid popularity for its claimed 397B parameters and open-source MIT license, but faced immediate scrutiny from the AI development community. IplanRIO has since apologized, attributing the release to an operational error where they mistakenly uploaded a merged baseline version instead of their independently trained model, and plan to re-upload a properly distilled and post-trained version. AI

    Global AI 'Dark Horse Myth' Reverses Overnight, Accused of Shelling Alibaba's Qwen

    IMPACT Highlights the importance of transparency and proper attribution in the open-source AI community, potentially impacting trust in new model releases.

  2. Rio de Janeiro's 'Homegrown' AI Was Someone Else's Model With a New Name

    The City of Rio de Janeiro's IT agency, IplanRIO, claimed to have developed an original 397-billion-parameter AI model named Rio-3.5-Open-397B. This model reportedly outperformed Alibaba's Qwen 3.7 Plus on several coding and reasoning benchmarks, generating significant public attention. However, an investigation by Nex-AGI revealed that Rio-3.5-Open-397B was not an original creation but rather a direct merge of Nex-AGI's Nex-N2-Pro model and Alibaba's Qwen3.5-397B-A17B base, with the model even identifying itself as Nex-AGI's creation when prompted. AI

    IMPACT Exposes potential misrepresentation in AI model claims, urging greater scrutiny of "homegrown" AI developments.