Models/Nemotron 3 Nano Omni 30B A3B Reasoning BF16

Text / nvidia

Nemotron 3 Nano Omni 30B A3B Reasoning BF16

NVIDIA Nemotron 3 Nano Omni 30B A3B Reasoning BF16 is a multimodal large language model described in its official model card as a 31B-parameter (approximately 3B active per token) Mamba2-Transformer hybrid Mixture-of-Experts model. It accepts video, audio, image, and text inputs and produces text outputs. It supports a maximum context of 256k tokens, has a reasoning mode enabled by default that can be toggled via enable_thinking, and is released under the NVIDIA Open Model Agreement as an any-to-any pipeline model using safetensors. The card lists BF16, FP8, and NVFP4 formats and various GPU suggestions, including minimum BF16 single-GPU use with 1× H100 80GB. The model is positioned for enterprise multimodal understanding workflows such as Q&A, summarization, transcription, document intelligence, and GUI/agentic automation. No testing, performance benchmarks, or compatibility claims beyond the provided card are inferred here.

EstimatedRepository opened Apr 20, 2026Source checked 9/25/2026Version: e5e99324

LOCALRENTED GPUOPEN WEIGHTS

Freshness 50 out of 100. Repository age: 178 days old, Apr 20, 2026. Measures age, not quality.

Popularity 74 out of 100. 336,466 downloads in 30 days; 439 lifetime likes. Checked Oct 15, 2026. Measures adoption, not quality.

What do these scores mean?

Two ways to compare models. Neither score measures answer quality, benchmark performance, or whether a model fits your machine.

Freshness · release age

A new release starts at 100. The score halves every 180 days: 100 × 2^(−age in days / 180). When the release date is unavailable, a repository date is a fallback—not proof of when the model was released. Missing dates stay unscored.

Check the date source

Popularity · adoption

Downloads over the past 30 days carry 80% of the score; lifetime likes carry 20%. Both use a logarithmic scale, capped at 10 million downloads and 100,000 likes. If likes are unavailable, downloads carry all the weight. Missing download statistics stay unscored.

round(80 × min(1, log10(1 + downloads) / 7) + 20 × min(1, log10(1 + likes) / 5))

Check the statistics source

Scores as of Oct 15, 2026. Statistics checked Oct 15, 2026. Statistics are marked for refresh after 72 hours.Full scoring method

WORKFLOWS FOR THIS MODEL 1

What do you want to do?

Choose a task. Get the software, files, settings, and steps to run it.

MAKE IT PERSONALChoose your hardware

Private chat

Chat privately with Nemotron 3 Nano Omni

Use a source-linked Q4_K_M conversion of NVIDIA Nemotron 3 Nano Omni in Ollama for a reviewable local text chat. The publisher BF16 and full multimodal serving routes remain linked separately.

Freshness 50 out of 100. Repository age: 178 days old, Apr 20, 2026. Measures age, not quality.

Ollama

1 packagemacOS / Windows / Linux

Open chat guide

Open a guide for its software, hardware requirements, and setup steps.

At a glance

Parameters: 33.02B
Architecture: NemotronH_Nano_Omni_Reasoning_V3
License: nvidia-open-model-agreement
Disk space: 61.5 GB
Software: vLLM

View the model source

MY HARDWARE

Will it run on your machine?

Save your machine to see a personalized rating and its reasoning.

Add my hardware

Best for

Enterprise multimodal Q&A over video, audio, image, and text inputs
Summarization and analysis of rich enterprise content such as meeting recordings and training videos
Document intelligence including OCR, chart understanding, and long-document processing
Automatic speech recognition and speech transcription workflows
GUI and agentic workflows such as browser agents, email agents, and incident management

Tradeoffs

Text-only output despite multimodal input; the model outputs text, not audio, video, or image modalities
The model card does not provide benchmark scores, accuracy metrics, or independent evaluation results
Use is governed by the NVIDIA Open Model Agreement and may be subject to its terms; no other license rights should be assumed
The BF16 variant is associated with large memory needs; the card lists 1× H100 80GB as minimum single-GPU for BF16, but actual compatibility depends on the specific deployment configuration and is not independently verified here
No inference of missing facts, performance characteristics, compatibility, or numerical hardware requirements is made beyond what is explicitly stated in the provided model card

Ways to run it

vLLM · Linux

Repository-specific command in publisher documentation; confirm dependencies and hardware in the source.

vllm serve nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4  --host 0.0.0.0  --max-model-len 131072  --tensor-parallel-size 1  --trust-remote-code  --video-pruning-rate 0.5  --max-num-seqs 384  --allowed-local-media-path /  --media-io-kwargs '{"video": {"fps": 2, "num_frames": 256}}'  --reasoning-parser nemotron_v3  --enable-auto-tool-choice  --tool-call-parser qwen3_coder \

Official instructions

Explore its uses

SAFETENSORSANY-TO-ANY

Keep exploring

Text

Prism ML

Bonsai 2 · 27B

A compressed 27B-class reasoning model with publisher-provided GGUF packs and a dedicated llama.cpp fork for CUDA, Metal, and CPU.

Freshness 90 out of 100. Release age: 28 days old, Sep 17, 2026. Measures age, not quality.

MODEL SIZE27.36B

LOCALRENTED GPU

Released Sep 17, 2026

Text

DeepSeek

DeepSeek V4.1 Flash

A multimodal reasoning model with a compressed key-value cache, published as open weights.

Freshness 87 out of 100. Repository age: 35 days old, Sep 10, 2026. Measures age, not quality.

Arena Open Source Text: rank 5 of 225. Ranked model: deepseek-v4.1-flash-max. Export dated Oct 13, 2026. Source checked Oct 15, 2026. Human preference ranking, not hardware suitability.+3

LOCALRENTED GPU

Repository opened Sep 10, 2026

Text

Qwen

Qwen3.8

Explore 4 versions of Qwen3.8. Compare model sizes, packages, hardware requirements, and available setups.

4 versionsCompare sizes & formats

LOCALRENTED GPU

Compare versions →