InferX Catalog | NVIDIA-Nemotron-Nano-12B-v2-VL-FP8

NVIDIA-Nemotron-Nano-12B-v2-VL-FP8

NVIDIA-Nemotron-Nano-12B-v2-VL-FP8 is a cutting-edge multimodal model released by NVIDIA in late 2025. It is specifically engineered for high-throughput, low-latency applications like document intelligence and long-form video understanding.

nvidia multimodal image2text

Log in to deploy

Metadata

Provider

nvidia

Modality

multimodal

API type

image2text

Source

huggingface / nvidia/NVIDIA-Nemotron-Nano-12B-v2-VL-FP8

Created

2026-04-12 14:28:21 UTC

Updated

2026-04-13 13:46:11 UTC

Catalog version

2

Visibility

Published

Specifications

Parameters

—

MoE

No

Max model length

10000

Image

vllm/vllm-openai:v0.16.0

Default Deploy Config

GPU count

1

vRAM

26000 MB

Summary

1xGPU 26000 MB

Recommended Use Cases

—

Model Spec

{
    "image": "vllm/vllm-openai:v0.16.0",
    "commands": [
        "--model",
        "nvidia/NVIDIA-Nemotron-Nano-12B-v2-VL-FP8",
        "--trust-remote-code",
        "--max-model-len",
        "10000",
        "--tensor-parallel-size=1"
    ],
    "resources": {
        "GPU": {
            "Count": 1,
            "vRam": 26000
        }
    },
    "envs": [],
    "policy": {
        "Obj": {
            "min_replica": 0,
            "max_replica": 1,
            "standby_per_node": 1,
            "parallel": 50,
            "queue_len": 100,
            "queue_timeout": 30.0,
            "scalein_timeout": 1.0,
            "scaleout_policy": {
                "WaitQueueRatio": {
                    "wait_ratio": 0.1
                }
            },
            "runtime_config": {
                "graph_sync": false
            }
        }
    },
    "sample_query": {
        "body": {
            "max_tokens": "200",
            "temperature": "0"
        },
        "path": "v1/chat/completions",
        "prompt": "What is in this image?",
        "apiType": "image2text",
        "dataUrl": "https://www.ilankelman.org/stopsigns/australia.jpg",
        "prompts": [],
        "loadingTimeout": 90
    }
}