NVIDIA GTC Berlin open-model drop: Nemotron 3.5 Super VL, Retriever, RFD4-Proteina
AllThingsApx · x · 2026-10-21
At GTC Berlin NVIDIA released a batch of open models: Nemotron 3.5 Super VL (already on Hugging Face), Nemotron Retriever covering embedding, reranking and doc parsing, and BioNeMo RFD4-Proteina, which designs proteins 13x faster than RFdiffusion3 in NVIDIA's benchmark. Aimed at European teams building agents, robotics, protein design and quantum research on their own infrastructure.
Related event: NVIDIA Unveils Nemotron 3.5 and Open-Weight Model Family at GTC Berlin(7 posts)→
More from Models
- OpenAI models' task horizon just 1/12 of METR trend — about 14 months behind, says Toby Ord — tobyordoxford · 2026-10-22
- InternScience's Sage-A1.5 trends on Hugging Face: a Qwen3.5-MoE vision agentic model — InternScience · 2026-10-22
- Mirendil-Base: one researcher pretrained a 100B+ coding base model in 3.5 months for under $10M — bneyshabur · 2026-10-22
- OpenAI's own report shows models' 80% time horizon on R&D is just 15 minutes — tobyordoxford · 2026-10-22
- Jensen Huang at GTC: Post-Training Nemotron Turns Open Models Into Frontier Capability — saranormous · 2026-10-22
- OpenAI and Apollo Research find o3 can detect when it's being safety-tested — alex_verem · 2026-10-22