/
← Voltar ao feed

Nvidia releases Nemotron 3.5 Super VL open-weight model as US open models gain ground after Hugging Face breach

1 fonte(s) · 1 atualização(ões) · publicado 15/10/2026 12:00 · coletado há 3 dias

Nvidia has released Nemotron 3.5 Super VL, a 120-billion-parameter open-weight model aimed at agentic workloads, together with two edge models, Nemotron 3.5 Bolt 74B-A9B 24GB and Nemotron 3.5 Bolt E4B 4GB. Nvidia claims it is 5x faster than the next-best model in its class, based on Artificial Analysis benchmarks that ranked InclusionAI's Ling 3.0 Flash and DeepSeek's V4.1 Flash second and third. The article places the release in a wider shift toward Western open-weight models, citing Mistral's Le Chonk and Reflection AI's Beam, against the earlier lead of Chinese labs such as DeepSeek and Moonshot.ai. It also says the OpenAI Hugging Face breach dispelled the notion that only open models carry security risk, and that cybersecurity guidance against foreign technology could favour US-made open models. The breach is mentioned only as background; the article gives no incident details, and this is product and industry news with no direct threat to defenders.

Linha do tempo

  1. zdnet.com ↗ · pub. 15/10/2026 12:00 · coletado 19/10 17:55 primeiro registro

    Nvidia releases Nemotron 3.5 Super VL open-weight model as US open models gain ground after Hugging Face breach

    ler a análise desta fonte

    Nvidia has released Nemotron 3.5 Super VL, a 120-billion-parameter open-weight model aimed at agentic workloads, together with two edge models, Nemotron 3.5 Bolt 74B-A9B 24GB and Nemotron 3.5 Bolt E4B 4GB. Nvidia claims it is 5x faster than the next-best model in its class, based on Artificial Analysis benchmarks that ranked InclusionAI's Ling 3.0 Flash and DeepSeek's V4.1 Flash second and third. The article places the release in a wider shift toward Western open-weight models, citing Mistral's Le Chonk and Reflection AI's Beam, against the earlier lead of Chinese labs such as DeepSeek and Moonshot.ai. It also says the OpenAI Hugging Face breach dispelled the notion that only open models carry security risk, and that cybersecurity guidance against foreign technology could favour US-made open models. The breach is mentioned only as background; the article gives no incident details, and this is product and industry news with no direct threat to defenders.