NVIDIA's open multimodal model (120B total, 12B active parameters) for image, video and document understanding.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Nemotron 3.5 Super VL is a vision-language model from NVIDIA with about 120B total parameters and 12B active per token. It takes images, video and documents as input and produces text. According to a creator review, it reaches over 600 tokens per second, and the checkpoint is about 75 GB.
The same review describes it as not the most capable open model on general intelligence, but notes that NVIDIA has released the training recipe, which makes it useful for researchers and teams that want to reproduce or adapt the model. The official NVIDIA pages fetched here did not describe the model, so details such as license, exact release date and benchmarks are unverified.
Access is presumably through NVIDIA's developer channels (API catalog, Hugging Face), but no direct URLs were confirmed in the provided material.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Open multimodal model (120B, 12B active) that understands images, video and documents, with over 600 tokens per second. Not the most intelligent open model, but training recipe is released; about 75 GB.
120B open-weight model scoring 26 on Artificial Analysis, below DeepSeek, Qwen and GLM. Its selling point is high output speed.
Closest neighbours by meaning, not just by category.
Model · Nvidia
NVIDIA's family of open models for agentic AI: reasoning, vision, speech, retrieval and safety models.
Model · NVIDIA
NVIDIA's open-weight ~550B mixture-of-experts LLM with a hybrid Mamba-transformer design, aimed at agentic workloads.
Model · NVIDIA
NVIDIA's open-weight mixture-of-experts speech model for offline transcription in 40 languages, up to one hour of audio.
Model · NVIDIA
NVIDIA's open-weight 30B mixture-of-experts model with 1M-token context, aimed at long-running agent workloads.
Model · NVIDIA
NVIDIA family of multilingual text embedding models (8B and 1B) for retrieval, RAG and coding agents.
Model · Nvidia
Nvidia research model that understands and generates images and video directly in pixel space, with no VAE or vision encoder.
