NVIDIA 新模型或于下周 GTC 发布，疑似 Nemotron 3.5 Super VL
NVIDIA 🔥: A new model is set to arrive next week, possibly at the upcoming NVIDIA GTC event on Wednesday the 21st.
NVIDIA 一款新模型预计下周发布，可能亮相于周三 21 日的 GTC 活动。此前在 NVIDIA 文档中出现的 Nemotron 3.5 Super VL 采用 Nemotron-H 混合语言解码器、RADIO 视觉编码器、视觉投影器和独立时序视频嵌入器，复用 Nemotron Omni 图像/视频栈但不含音频编码器，并保留共享的 Multi-Token Prediction（MTP）权重。引用信息还提到该模型为 120B 总参数、12B 激活参数的混合 Mamba-Transformer MoE，支持图像和视频输入，并有纯文本版本。
查看原帖NVIDIA's next Nemotron showed up in its own docs: Nemotron 3.5 Super VL. • 120B total, 12B active (hybrid Mamba-Transformer MoE) • Image and video input • Text-only version too NVIDIA started pulling the pages today, the date a partner issue listed as its target release. https://t.co/YGiqsKGaiP
正文
当前显示原文，中文全文译文尚未就绪。
NVIDIA 🔥: A new model is set to arrive next week, possibly at the upcoming NVIDIA GTC event on Wednesday the 21st.
Are we getting the recently spotted Nemotron 3.5 Super VL, or something else on top of it?
> "Nemotron 3.5 Super VL combines a Nemotron-H hybrid language decoder with a RADIO vision encoder, a vision projector, and a separate temporal video embedder. It reuses the Nemotron Omni image/video stack without an audio encoder and retains shared Multi-Token Prediction (MTP) weights."
Monitoring 👀