All tasks / Models
Nemotron-3.5-Embed-8B
NVIDIA's open-weights embedding models for text and visual-document retrieval, in 8B and 1B sizes.
- NVIDIA
- Open weights
- Released Oct 2026
Where it ranks
- –Embeddings & retrievalModels · Local · no result on this task’s benchmarks yet
Facts
- Vendor
- NVIDIA
- Family
- Nemotron Embed
- Released
- Oct 2026
- Weights
- Open: download and run it yourself
- License
- OpenMDW-1.1
- Parameters
- 8.38B
- GPU memory to run locally
- ~17 GB (Derived: 8.38B parameters x 2 bytes = ~16.8 GB of bf16 weights; a 1B size also exists.)
- Access
- Free (open weights)
- Last verified
- 19 Oct 2026
Benchmark results
No public benchmark results recorded yet.
Where you can use it
No tracked tool lists this model yet; as open weights it runs in local runtimes.
Earlier versions
- Nemotron-3-Embed-8B · Jul 2026
- llama-embed-nemotron-8b · Oct 2025