Nemotron 3 Nano 30B A3B
NVIDIAfamily · Nemotron 3huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B
NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems. The model is fully...
Updated 4 h ago · first seen 11 Sept 2026
model_01M294WW73MBCT953NW5VQH309
Overview
Identity
- Canonical model
- Yesidentity confidence: mediumOne row per real model release. Artifacts (checkpoints, quantisations, conversions) and folded evaluation variants point here.
- Official checkpoints
- official_checkpoints = hf_repo identifiers carried by the model itself; artifacts are separate entities pointing here through canonical_id.
- Artifacts
- 1 quantization0 official · 1 third-partySeparate entities (checkpoint · quantization · conversion · packaging) pointing to this model through canonical_id.
- Provider deployments
- 2
- API aliases
- nvidia-nemotron-3-nano-30b-a3bnvidia-nemotron-3-nano-30b-a3b-reasoningnvidia/nemotron-3-nano-30b-a3bIdentifiers under which providers and evaluators refer to this model.
- Folded evaluation variants
- 1Effort / thinking variants (…-high, …-non-reasoning) are result configurations of this model, not separate models. Their old URLs redirect here.
Openness
Restricted weights— weights downloadable under NVIDIA-Open-Model; commercial use allowed; redistribution allowed; derivatives allowed; 4 dimensions unknown.
Weights downloadable, but the licence restricts commercial use, hosting, derivatives or field of use (community, research and RAIL licences).
Weights
Yes
Inference code
—
Training code
—
Training data
—
Dataset
—
Commercial use
Yes
Redistribution
Yes
Derivatives
Yes
Licence: NVIDIA Open Model License (community)
dimensions marked null are unknown, not false
Key facts
- Release date
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Model card
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Openrouter id
Source:OpenRouter public model & pricing listingT2observed 5 d agomedium
Architecture
- Architecture
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Model type
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Parameters
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Active parameters
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Weights dtype
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- File size
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Library name
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Pipeline tag
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
- Hugging Face repo
Source:OpenRouter public model & pricing listingT2observed 5 d agomedium
Capabilities
Modalities
- Modalities
- text
- Input
- text
- Output
- text
Capabilities
Tool calling
Yes
OpenRouter public model & pricing listing · T2
Structured output
Yes
OpenRouter public model & pricing listing · T2
Reasoning
Yes
OpenRouter public model & pricing listing · T2
Vision
Unavailable
Audio
Unavailable
Fine-tuning available
Unavailable
- Context window
Source:OpenRouter public model & pricing listingT2observed 5 d agomedium
- Max output
Source:OpenRouter public model & pricing listingT2observed 5 d agomedium
- Languages
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 5 d agomedium
Benchmarks30
Compare with another model →Comparable same task and conditions · Partially comparable same task, conditions differ (effort, temperature, judge) · Not comparable different variant or metric
Current rows only, grouped by benchmark → canonical metric → comparability group (task configuration). Effort variants folded into this model appear as rows of the same group. 30 current rows in total. “vs leader” compares with the current leader of the benchmark's primary group only; other groups are not directly comparable. Comparability rules →
Providers & Pricing2
All offers in the price terminal →USD per 1M tokens as published by each provider; native units (per-request fees, flex/priority tiers) are kept verbatim. Rows are append-only — every price change is kept in the history below. Cost of a workload →
Price history
Output price · USD / 1M tokens 2 providers
- NVIDIA NIM / build.nvidia.com
- OpenRouter
- OpenRouterfirst observed $0.2012 Sept 2026
- NVIDIA NIM / build.nvidia.comfirst observed $0.2011 Sept 2026
Input price · USD / 1M tokens 2 providers
- NVIDIA NIM / build.nvidia.com
- OpenRouter
- OpenRouterfirst observed $0.0512 Sept 2026
- NVIDIA NIM / build.nvidia.comfirst observed $0.0511 Sept 2026
Hardware fit37
Assumptions (7)
- Estimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead (or the observed artifact file size when one is recorded).
- bytes/param: 4bit = 0.5, 8bit = 1.0, fp16 = 2.0 (uniform quantization, no per-layer exceptions).
- KV cache: 2 × layers × kv_heads × head_dim × 2 bytes × context × batch when the architecture is known; otherwise 0.5 GB per 8 192 tokens (× batch), independent of architecture (GQA/MLA models need less).
- A model 'fits' when the estimate is at most the device memory minus 2 GB reserved for the OS and framework.
- Mixture-of-experts models are estimated on total parameters (all experts must be resident); active parameters are ignored.
- Device memory uses the largest configuration when several are listed (e.g. Apple silicon tiers).
- Multi-GPU: device memories are summed; interconnect bandwidth, tensor-parallel replication and pipeline bubbles are not modelled.
Lineage
Open in Graph →Versions & Artifacts1
Version history
Context window2 changes
11 Sept 2026→11 Sept 2026→12 Sept 2026current
License1 change
11 Sept 2026→12 Sept 2026current
Openness1 change
11 Sept 2026→12 Sept 2026current
Max outputfirst observation only
11 Sept 2026current
Parametersfirst observation only
11 Sept 2026current
Each hop is a claim: click a value for its source, tier and observation time. Nothing is overwritten — a new observation closes the previous claim.
Artifacts 1
quantization 1
- nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4NVIDIA · BF16/F32/F8_E4M3/U819.4 GB
Papers2
- arXiv:2512.20856Active35
- arXiv:2512.20848Active35
Datasets16
- 35
- nvidia/Nemotron-CC-v2.1Active35
- nvidia/Nemotron-Math-v2Active35
- 35
- nvidia/Nemotron-CC-v2Active35
- 35
- 35
- 35
Timeline22
Full timeline →Nemotron 3 Nano 30B A3B: openness changed from open-weights to restricted-weights
Opennessopen-weights→restricted-weightshuggingfaceNemotron 3 Nano 30B A3B: license changed from Other to NVIDIA-Open-Model
LicenseOther→NVIDIA-Open-ModelhuggingfaceOpenRouter lists Nemotron 3 Nano 30B A3B at $0.05 in / $0.2 out per 1M tokens
openrouterNemotron 3 Nano 30B A3B scores 12.12% on Terminal-Bench
artificial_analysisNemotron 3 Nano 30B A3B scores 25.44% on τ²-bench
artificial_analysisNemotron 3 Nano 30B A3B scores 37.48% on IFBench
artificial_analysisNemotron 3 Nano 30B A3B scores 4.59% on Humanity's Last Exam
artificial_analysisNemotron 3 Nano 30B A3B scores 39.9% on GPQA Diamond
artificial_analysisNemotron 3 Nano 30B A3B scores 6.85 on Artificial Analysis Intelligence Index
artificial_analysisNemotron 3 Nano 30B A3B scores 13.64% on Terminal-Bench
artificial_analysisNemotron 3 Nano 30B A3B scores 6.74% on Terminal-Bench
artificial_analysisNemotron 3 Nano 30B A3B scores 0% on Terminal-Bench
artificial_analysisNemotron 3 Nano 30B A3B scores 40.94% on τ²-bench
artificial_analysisNemotron 3 Nano 30B A3B scores 71.09% on IFBench
artificial_analysisNemotron 3 Nano 30B A3B scores 30.56% on SciCode
artificial_analysisNemotron 3 Nano 30B A3B scores 11.4% on Humanity's Last Exam
artificial_analysisNemotron 3 Nano 30B A3B scores 75.66% on GPQA Diamond
artificial_analysisNemotron 3 Nano 30B A3B scores 8.9 on Artificial Analysis Intelligence Index
artificial_analysisNemotron 3 Nano 30B A3B: reasoning changed from false to true
ReasoningNo→YesopenrouterNemotron 3 Nano 30B A3B: context length changed from 1000000 to 262144
Context window1M tokens→262.1K tokensopenrouterNemotron 3 Nano 30B A3B: release date changed from 2025-12-04 to 2025-12-14
Release date4 Dec 2025→14 Dec 2025openrouterNemotron 3 Nano 30B A3B: release date changed from 2025-12-14 to 2025-12-04
Release date14 Dec 2025→4 Dec 2025huggingface
Change history75
Viewing AI Atlas as of 16 Mar 2026 — attributes exactly as the atlas knew them on that day; later corrections are not shown.
Back to today →Nemotron 3 Nano 30B A3B was not yet in AI Atlas on 16 Mar 2026
Release daterelease_date6
Opennessopenness3
Licenselicense2
Architecturearchitecture1
Parametersparameter_count1
Active parametersactive_parameter_count1
Context windowcontext_length4
Max outputmax_output_tokens1
Modalitiesmodalities1
Input modalitiesmodalities_input1
Output modalitiesmodalities_output1
Languageslanguages1
File sizefile_size_gb1
Hugging Face repohf_repo1
Pipeline tagpipeline_tag1
Model cardmodel_card_url1
Downloadsmetric.downloads6
Likesmetric.likes4
Aa context windowaa_context_window1
Aa opennessaa_openness1
Accessaccess1
Artifact kindartifact_kind1
Commercial use allowedcommercial_use_allowed1
Datasetsdatasets1
Derivatives allowedderivatives_allowed1
Descriptiondescription1
Gatedgated1
Hf inference providershf_inference_providers1
Last modifiedlast_modified1
Library namelibrary_name1
License keylicense_key1
License namelicense_name1
License rawlicense_raw1
License urllicense_url1
Aa median output tokens per secondmetric.aa_median_output_tokens_per_second2
Downloads all timemetric.downloads_all_time6
Model typemodel_type1
Openrouter idopenrouter_id1
Openrouter listed atopenrouter_listed_at1
Reasoningreasoning4
Redistribution allowedredistribution_allowed1
Structured outputstructured_output1
Supported parameterssupported_parameters1
Tool callingtool_calling1
Weights availableweights_available1
Weights dtypeweights_dtype1
Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →
Provenance
Attributed facts
48
Source tiers
T248
Freshest observation
5 h ago
Conflicts
None
Source documents 6
Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.
Data quality (72/100) measures how well AI Atlas knows this entity — completeness, primary-source ratio, freshness, conflicts — never how good the model is. Methodology →