Nemotron 3 Nano 30B A3B
NVIDIAhuggingface.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 11 min ago · first seen 11 Sept 2026
model_01M294WW73MBCT953NW5VQH309
- Parameters
- 31.6B · 3B active
- T2 · 2 h ago
- Context
- 1M tokens
- T2 · 50 min ago
- Released
- 4 Dec 2025
- T2 · 11 min ago
- License
- other
- T2 · 2 h ago
Specification
- Release date
- 4 Dec 2025
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 11 min agomedium
- Openness
- open-weights
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- License
- other
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Architecture
- NemotronHForCausalLM
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Parameters
- 31.6B
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Active parameters
- 3B
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Context window
- 1M tokens
Source:Artificial AnalysisT2observed 50 min agomedium
- Max output
- 235.9K tokens
Source:OpenRouter public model & pricing listingT2observed 2 h agomedium
- Modalities
- text
Source:OpenRouter public model & pricing listingT2observed 2 h agomedium
- Input modalities
- text
Source:OpenRouter public model & pricing listingT2observed 2 h agomedium
- Output modalities
- text
Source:OpenRouter public model & pricing listingT2observed 2 h agomedium
- Languages
- en, es, fr, de, ja, it
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- File size
- 63.2 GB
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Hugging Face repo
- nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
Source:OpenRouter public model & pricing listingT2observed 2 h agomedium
- Pipeline tag
- text-generation
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Model card
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Downloads
- 804,613
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 11 min agomedium
- Likes
- 820
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 11 min agomedium
- Datasets
- nvidia/Nemotron-Pretraining-Code-v1, nvidia/Nemotron-CC-v2, nvidia/Nemotron-Pretraining-SFT-v1, nvidia/Nemotron-CC-Math-v1, nvidia/Nemotron-Pretraining-Code-v2, nvidia/Nemotron-Pretraining-Specialized-v1, nvidia/Nemotron-CC-v2.1, nvidia/Nemotron-CC-Code-v1, nvidia/Nemotron-Pretraining-Dataset-sample, nvidia/Nemotron-Competitive-Programming-v1, nvidia/Nemotron-Math-v2, nvidia/Nemotron-Agentic-v1, nvidia/Nemotron-Math-Proofs-v1, nvidia/Nemotron-Instruction-Following-Chat-v1, nvidia/Nemotron-Science-v1, nvidia/Nemotron-3-Nano-RL-Training-Blend
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Gated
- No
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Hf inference providers
- featherless-ai
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Last modified
- 2026-08-24T23:50:04+00:00
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Library name
- transformers
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- License name
- nvidia-nemotron-open-model-license
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
- License url
- https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-nemotron-open-model-license/
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Aa median output tokens per second
- 151.2
Source:Artificial AnalysisT2observed 50 min agomedium
- Downloads all time
- 9,342,427
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 11 min agomedium
- Model type
- nemotron_h
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Openrouter id
- nvidia/nemotron-3-nano-30b-a3b
Source:OpenRouter public model & pricing listingT2observed 2 h agomedium
- Supported parameters
- frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p
Source:OpenRouter public model & pricing listingT2observed 2 h agomedium
- Tags
- transformers, safetensors, nemotron_h, text-generation, nvidia, pytorch, custom_code, en, es, fr, de, ja, it
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
- Weights dtype
- BF16, F32
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
Each value shows its source, tier and observation time. Conflicting claims are kept side by side and flagged — never averaged. How AI Atlas records facts →
Provenance
Attributed facts
36
Source tiers
T236
Freshest observation
11 min ago
Conflicts
None
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
No
Artificial Analysis · T2
Vision
Unavailable
Audio
Unavailable
Fine-tuning available
Unavailable
- Context window
- 1M tokens
Source:Artificial AnalysisT2observed 50 min agomedium
- Max output
- 235.9K tokens
Source:OpenRouter public model & pricing listingT2observed 2 h agomedium
- Languages
- en, es, fr, de, ja, it
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 2 h agomedium
| Benchmark | Score | Metric | Config | Evaluated | Source |
|---|---|---|---|---|---|
| τ²-bench | 25.44% | pass^1 | aa_slug=nvidia-nemotron-3-nano-30b-a3b · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
| Artificial Analysis Intelligence Index | 6.85 | index | aa_slug=nvidia-nemotron-3-nano-30b-a3b · version=4.3 · estimated=true | — | artificialanalysis.aiT2 |
| GPQA | 39.9% | accuracy | aa_slug=nvidia-nemotron-3-nano-30b-a3b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
| Humanity's Last Exam | 4.59% | accuracy | aa_slug=nvidia-nemotron-3-nano-30b-a3b · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
| IFBench | 37.48% | accuracy | aa_slug=nvidia-nemotron-3-nano-30b-a3b · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
| Terminal-Bench | 12.12% | accuracy | aa_slug=nvidia-nemotron-3-nano-30b-a3b · variant=hard · evaluator=Artificial Analysis · index_version=4.3 | — | artificialanalysis.aiT2 |
Scores are reported as published, with their evaluation configuration (harness, prompting, judge). Results with different configs are not directly comparable — see methodology.
Current prices
| Provider | Input / 1M | Output / 1M | Cached in | Batch in / out | Context | Observed | Source |
|---|---|---|---|---|---|---|---|
| NVIDIA NIM / build.nvidia.comnvidia/nemotron-3-nano-30b-a3b | $0.05 | $0.20 | $0.03 | — | 262.1K | 55 min ago | openrouter.aiT2 |
USD per 1M tokens as published by each provider (USD). Rows are append-only: every change is kept in the history below.
Available through 1
Price history
Input price · USD / 1M tokens
Output price · USD / 1M tokens
- NVIDIA NIM / build.nvidia.com
Memory need = bytes per parameter (4-bit ≈ 0.5 × 1.15 overhead, 8-bit 1.0, fp16 2.0) + a KV-cache allowance. Not a measurement.
| Hardware | Quantization | Memory | Est. need | Fits |
|---|---|---|---|---|
| NVIDIA DGX B200 | 4bit | 1,440 GB | 18.7 GB | Yes |
| Apple M3 Ultra | 4bit | — | 18.7 GB | Yes |
| Mac Studio (Apple M5 Ultra) | 4bit | — | 18.7 GB | Yes |
| AMD Instinct MI325X | 4bit | 256 GB | 18.7 GB | Yes |
| AMD Instinct MI300X | 4bit | 192 GB | 18.7 GB | Yes |
| Apple M2 Ultra | 4bit | — | 18.7 GB | Yes |
| NVIDIA B200 | 4bit | 180 GB | 18.7 GB | Yes |
| NVIDIA H200 | 4bit | 141 GB | 18.7 GB | Yes |
| NVIDIA H200 NVL | 4bit | 141 GB | 18.7 GB | Yes |
| Apple M1 Ultra | 4bit | — | 18.7 GB | Yes |
| Apple M3 Max | 4bit | — | 18.7 GB | Yes |
| Apple M4 Max | 4bit | — | 18.7 GB | Yes |
| Mac Studio (Apple M5 Max) | 4bit | — | 18.7 GB | Yes |
| MacBook Pro (Apple M5 Max) | 4bit | — | 18.7 GB | Yes |
| NVIDIA DGX Spark | 4bit | 128 GB | 18.7 GB | Yes |
| Apple M2 Max | 4bit | — | 18.7 GB | Yes |
| NVIDIA H100 NVL | 4bit | 94 GB | 18.7 GB | Yes |
| NVIDIA A100 80GB | 4bit | 80 GB | 18.7 GB | Yes |
| NVIDIA H100 SXM | 4bit | 80 GB | 18.7 GB | Yes |
| Apple M1 Max | 4bit | — | 18.7 GB | Yes |
| Apple M4 Pro | 4bit | — | 18.7 GB | Yes |
| Mac mini (Apple M5 Pro) | 4bit | — | 18.7 GB | Yes |
| MacBook Pro (Apple M5 Pro) | 4bit | — | 18.7 GB | Yes |
| Apple M3 Pro | 4bit | — | 18.7 GB | Yes |
| Apple M1 Pro | 4bit | — | 18.7 GB | Yes |
| Apple M2 Pro | 4bit | — | 18.7 GB | Yes |
| Apple M4 | 4bit | — | 18.7 GB | Yes |
| iMac (Apple M4) | 4bit | — | 18.7 GB | Yes |
| Mac mini (Apple M6) | 4bit | — | 18.7 GB | Yes |
| MacBook Air (Apple M5) | 4bit | — | 18.7 GB | Yes |
| MacBook Pro (Apple M5) | 4bit | — | 18.7 GB | Yes |
| NVIDIA GeForce RTX 5090 | 4bit | 32 GB | 18.7 GB | Yes |
| Apple M2 | 4bit | — | 18.7 GB | Yes |
| Apple M3 | 4bit | — | 18.7 GB | Yes |
| NVIDIA GeForce RTX 3090 | 4bit | 24 GB | 18.7 GB | Yes |
| NVIDIA GeForce RTX 4090 | 4bit | 24 GB | 18.7 GB | Yes |
| Apple M1 | 4bit | — | 18.7 GB | No |
No lineage recorded
Papers 2
- arXiv:2512.20856Active35
- arXiv:2512.20848Active35
Repositories 0
No repositories linked yet.
As of
Rewind the record: see this entity's attributes exactly as AI Atlas knew them on a given day.
Claim history · Gated
Gatedgated1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| No | → current | current | Hugging Face Hub (public pages, model cards, papers)T2 | medium | deterministic |
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 →
Nemotron 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
artificial_analysisNemotron 3 Nano 30B A3B scores 6.85 on Artificial Analysis Intelligence Index
artificial_analysisNemotron 3 Nano 30B A3B: context length changed from 262144 to 1000000
Context window262.1K tokens→1M tokensartificial_analysisNemotron 3 Nano 30B A3B: reasoning changed from true to false
ReasoningYes→Noartificial_analysisNemotron 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 2025huggingfaceNVIDIA NIM / build.nvidia.com lists Nemotron 3 Nano 30B A3B at $0.05 in / $0.2 out per 1M tokens
openrouter
| Source | Document | Type | Tier | Last observed | Snapshots |
|---|---|---|---|---|---|
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | model_page | T2· Quality secondary | 11 min ago | 2 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/models?author=nvidia&p=0&sort=downloads | listing | T2· Quality secondary | 39 min ago | 3 |
| Artificial Analysis | artificialanalysis.ai/leaderboards/models | leaderboard | T2· Quality secondary | 50 min ago | 1 |
| OpenRouter public model & pricing listing | openrouter.ai/api/v1/models | catalogue | T2· Quality secondary | 55 min ago | 2 |
| Hugging Face Hub (public pages, model cards, papers) | huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16/raw/main/README.md | model_card | T2· Quality secondary | 1 h ago | 1 |
Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.