[Models](/models)/[Captioning](/models?category=captioning)/nvidia/Cosmos3-Nano

CosmosScene CaptioningOther

# Cosmos3-Nano

by nvidia

16B omni model with text, image, video, audio, action generation, and video reasoner input

36.7Kdl/month

16Bparams

[HuggingFace](https://huggingface.co/nvidia/Cosmos3-Nano)[Run it on your own data, free](https://studio.mixpeek.com/namespaces/create?mode=managed&utm_source=homepage&utm_medium=model_page&utm_campaign=mvs_onramp)

Identifiers

Model ID

`nvidia/Cosmos3-Nano`

## Deploy Cosmos3-Nano

Single-tenant

Mixpeek has no managed extractor for this model. On a single-tenant deployment you upload the weights and a custom plugin serves them next to the rest of your pipeline.

[Read the customer-hosted guide](https://mixpeek.com/docs/customer-hosted/overview) [Talk to us about single-tenant](/demo)

## Overview

Cosmos3-Nano is a compact member of NVIDIA's Cosmos3 family. The model card describes generator inputs across text, image, video with or without audio, and action trajectory, plus a reasoner path that accepts text, text plus image, and text plus video, then returns text. That makes it relevant to agent perception work where a system needs to inspect or reason over a short video candidate.

On Mixpeek, Cosmos3-Nano is most useful after retrieval has selected a small set of clips. Store timeline metadata and keyframe embeddings first, then run a video reasoning pass to extract events, object interactions, or natural-language answers tied back to the source clip.

## Architecture

Cosmos3 omni model with generator and reasoner interfaces. The reasoner supports text, text plus image, and text plus video input with text output. The model card recommends video reasoner input around 4 fps and supports long-context inputs up to 256K tokens.

## Mixpeek SDK Integration

IngestRetrieve

// No extractor parameter takes a Hugging Face model id (checked against
// GET /v1/discovery/extractors, which returns 13), so Cosmos3-Nano runs
// on your side and the output is upserted through POST
// /v1/namespaces/{namespace\_id}/documents/upsert. On Enterprise the other
// path is to upload the weights instead: POST /v1/namespaces/{id}/models
// accepts the huggingface format and a custom plugin loads them.
const res = await fetch(
  "https://api.mixpeek.com/v1/namespaces/ns\_your\_namespace/documents/upsert",
  {
    method: "POST",
    headers: {
      Authorization: "Bearer API\_KEY",
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      collection\_id: "col\_your\_collection",
      documents: \[
        {
          document\_id: "asset-00412",
          // The model produces text, so it lands in payload. Give the
          // collection a text vector index and embed that text to make it
          // searchable rather than only filterable.
          payload: { extracted\_text: modelOutput, source\_key: "archive/2026/asset-00412" },
          vectors: { "multimodal-embedding": embeddingOfModelOutput },
        },
      \],
    }),
  },
);

// Managed alternative, if this exact model is not the requirement:
// universal\_extractor@v1 runs google/gemini-embedding-2
// (3072-d) over a bucket, with no inference of your own.

## Capabilities

*   Video reasoner input for short retrieved clips
*   Text and image conditioning for multimodal inspection
*   Video, audio, and action generation interfaces for simulation workflows
*   Long-context text handling around video evidence

## Use Cases on Mixpeek

Ask an agent to explain what happens in a retrieved video clip

Extract event descriptions from short candidate clips before reranking

Build visual inspection tools that reason over video, not only single frames

Prototype synthetic video or action data around agent perception evals

## Performance

Input SizeText, image, video, optional audio, and action inputs

GPU LatencyVideo length and output length dependent

GPU ThroughputBatch dependent

GPU Memory16B omni deployment class

Use on retrieved clips or sampled windows rather than every raw frame

## Common Pipeline Companions

[

NemoStation/Marlin-2B

Dense timestamped captions before detailed reasoning

](/model/NemoStation/Marlin-2B)[

google/videoprism-base-f16r288

First-stage video embedding retrieval

](/model/google/videoprism-base-f16r288)[

BAAI/bge-m3

Embed generated event summaries for search

](/model/BAAI/bge-m3)

## Explore on Mixpeek

[

More Captioning models

Compare alternatives in this category

](/models?category=captioning)[

Best Document AI Platforms

Hand-picked tools & platforms compared

](/curated-lists/best-document-ai-platforms)[

OCR & Document AI Internals

Deep-dive technical guide

](/guides/ocr-document-ai-internals)[

Feature Extractors

See how Mixpeek runs models as extractors

](/extractors)[

Mixpeek Vector Store

Store & search embeddings at scale

](/mvs)[

Pricing

Usage-based pricing for pipelines

](/pricing)[

All Curated Lists

Compare models, APIs & infrastructure

](/curated-lists)

### Specification

FrameworkCosmos

Organizationnvidia

FeatureScene Captioning

Outputtext

Modalitiesvideo, image

RetrieverSemantic Search

Parameters16B

LicenseOther

Downloads/mo36.7K

[

### Research Paper

Cosmos3-Nano model card

arxiv.org](https://huggingface.co/nvidia/Cosmos3-Nano)

### Build a pipeline with Cosmos3-Nano

Add this model to a processing pipeline alongside other extractors. Combine with retrieval stages for end-to-end search.

[Run it on your own data, free](https://studio.mixpeek.com/namespaces/create?mode=managed&utm_source=homepage&utm_medium=model_page&utm_campaign=mvs_onramp)

### Alternative Models

[

Salesforce/blip2-opt-2.7b

Scene Captioning

](/model/Salesforce/blip2-opt-2.7b)[

microsoft/Florence-2-large

Scene Captioning

](/model/microsoft/Florence-2-large)[

google/paligemma2-3b-mix-448

Scene Captioning

](/model/google/paligemma2-3b-mix-448)[

Qwen/Qwen3-VL-8B-Instruct

Scene Captioning

](/model/Qwen/Qwen3-VL-8B-Instruct)