NVIDIA
Nemotron 3.5 Parse
NVIDIA's 0.6B-parameter open vision-language model with a 16K-token context, ready to serve on Runpod Serverless with vLLM or on a Pod.
Nemotron 3.5 Parse
Specifications
Provider
NVIDIA
Modality
Vision + language
Task
Text generation
Parameters
0.6B
Active parameters
Context
16K tokens
Precision
BF16
Weights
2
License
openmdw-1.1
Released
2026-09-02
Endpoint ID
Weights on Hugging Face
nvidia/Nemotron-3.5-Parse
Frequently asked questions
How do I deploy Nemotron 3.5 Parse on Runpod?
Deploy the vLLM worker from the Runpod Hub as a Serverless endpoint and set MODEL_NAME to nvidia/Nemotron-3.5-Parse. The endpoint is OpenAI-compatible and scales to zero when idle. You can also run vLLM on a Pod.
What GPU do I need to run Nemotron 3.5 Parse?
Start with RTX 4090 (24 GB VRAM). The weights take about 2 GB; longer contexts and more concurrent requests need extra memory for the KV cache.
How much does Nemotron 3.5 Parse cost on Runpod?
Serverless endpoints and Pods bill per second for the GPU you choose. A Serverless endpoint costs nothing while scaled to zero.
nvidia-nemotron-3-5-parse
nvidia/Nemotron-3.5-Parse
Vision + language
2026-09-02 17:00
RTX 4090