[whatmodel.sh](../../README.md) / Nemotron Nano 9B v2 # Nemotron Nano 9B v2 NVIDIA · released 2025-08 · 8.89B params · nvidia-open-model **score 55.4** ±0.7 · #270 of 732 · from 34 benchmark results across 5 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at ● marks Nemotron Nano 9B v2 among every model on that task or test, worst to best. 
task            vs the others          rank  counts  tests
agent tasks     ▁▁▃▆●▄▅▅▄▄▂▁   #189 of 313*    1.3%      1
coding          ▁▁▁▃▅▇●▇▅▄▃▁    #231 of 469     22%      7
chat & writing  ▁▁▁▃▆▄▆●▆▇▇▃    #187 of 402    3.7%      4
math            ▁▁▃▅█▆●█▄▂▁▁    #167 of 431     33%      5
agent tasks · #189 of 313 · 1.3% of its score Beats [o1 (2024-12-17)](o1-2024-12-17.md), [o3-mini](o3-mini.md), [DeepSeek R1 0528](deepseek-r1-0528.md) here, though they rank higher overall. Trails [SmolLM3 3B](smollm3-3b.md), [Jamba 1.5 Mini](jamba-1-5-mini.md), [Mellum2 12B-A2.5B Instruct](mellum2-12b-a2-5b-instruct.md) here, though they rank lower overall. 
test             result  vs the others in its run    difficulty        counts  why
BFCL v3 (maker)   66.9%  ▁▁▁▁▁▃▅█▆●▄▁ beats 132/171  medium · avg 58%    1.3%  blunt · at its level · small run · seve…
coding · #231 of 469 · 22% of its score Beats [Qwen3.5 4B](qwen3-5-4b.md), [Qwen3-VL 32B Instruct](qwen3-vl-32b-instruct.md), [Solar Open 100B (102B-A12B)](solar-open-100b.md) here, though they rank higher overall. Trails [QwQ-32B-Preview](qwq-32b-preview.md), [Ministral 3 3B Reasoning 2512](ministral-3-3b-reasoning.md), [GPT-4 (0613)](gpt-4-0613.md) here, though they rank lower overall. 
test                                result  vs the others in its run  difficulty            counts  why
LiveCodeBench Pro 2025Q2 (Medium)…    4.8%  ▃▃····●·█·▃▃ beats 2/7    very hard · avg 2.6%    8.4%  sharp · too hard for it
LiveCodeBench Pro 2025Q2 (Easy) (…   59.3%  ▄·▄····●▄▄█▄ beats 2/7    medium · avg 46%        5.5%  at its level
LiveCodeBench v6 (Aug 2024–May 20…   65.3%  ▄▄···●·▄▄█·▄ beats 2/7    medium · avg 58%        2.4%  at its level
MBPP+                                66.1%  ▄▄····▄·●··█ beats 3/5    medium · avg 61%        1.7%  at its level
LiveCodeBench v3                     83.4%  ▃▃·····▃·▃·● beats 5/6    medium · avg 64%        1.5%  typical
HumanEval+                           89.7%  ▂·······▂·▂● beats 3/6    easy · avg 78%          1.4%  sharp · noisy
LiveCodeBench (version unspecifie…   71.1%  ▁···▁▁▁█▇●▄▁ beats 31/38  medium · avg 46%        1.1%  at its level · small run
chat & writing · #187 of 402 · 3.7% of its score Beats [Seed-OSS 36B Instruct](seed-oss-36b.md), [Gemma 2 9B IT SimPO](gemma-2-9b-simpo.md), [Nemotron-Cascade-14B-Thinking](nemotron-cascade-14b-thinking.md) here, though they rank higher overall. Trails [LFM2.5 1.2B Instruct](lfm2-5-1-2b-instruct.md), [LFM2.5 1.2B Thinking](lfm2-5-1-2b-thinking.md), [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md) here, though they rank lower overall. 
test                                result  vs the others in its run  difficulty        counts  why
AlpacaEval 2 length-controlled       58.0%  █···█·█·●·██ beats 3/5    medium · avg 43%    2.6%  sharp · at its level
IFEval prompt-level strict (maker)   86.1%  ▁▁··▂··▂▇●▃▃ beats 16/26  easy · avg 77%      0.9%  small run
IFEval (aggregation unspecified)     86.0%  ▄▄·····●▄▄·▄ beats 3/6    easy · avg 79%      0.3%  noisy
IFEval instruction-level strict (…   90.3%  ▄▄····▄▄█▄·● beats 8/8                           —  not counted yet: no other model in the …
math · #167 of 431 · 33% of its score Beats [MiMo-V2.5 (310B-A15B)](mimo-v2-5.md), [MiniMax-M2.7](minimax-m2-7.md), [MiniMax-M2.1](minimax-m2-1.md) here, though they rank higher overall. Trails [Olmo 3 7B Think](olmo-3-7b-think.md) here, though they rank lower overall. 
test                         result  vs the others in its run  difficulty           counts  why
AIME 2025                     58.9%  ▄··▄····●▄▄▄ beats 3/6    medium · avg 50%        16%  sharp · at its level · several runs
AIME 2024                     72.1%  ▄·····▄··▄●█ beats 4/6    medium · avg 58%       9.0%  sharp · at its level · several runs
MATH (subset unspecified)     94.4%  ▃·····▃···●█ beats 2/6    very easy · avg 91%    3.2%  maxed out
OMEGA (math generalisation)   42.4%  ▄······█▄▄·● beats 5/6    hard · avg 31%         3.1%  at its level
MATH-500 (maker)              97.8%  ▁▃▂▃▄▇▅▇█●▂▁ beats 74/85  very easy · avg 93%    1.3%  maxed out · small run
other tests (12) 
test                                result  vs the others in its run    difficulty            counts  why
GPQA (subset unspecified)            56.2%  ▃······▃·▅·● beats 4/6      hard · avg 43%           12%  sharp · at its level · several runs
MMLU                                 84.3%  ▄···▄··█··●█ beats 4/6      easy · avg 76%          6.8%  several runs
BIG-Bench Hard                       86.2%  ▄·····▄·▄·●█ beats 3/6      easy · avg 77%          5.0%  sharp
GPQA Diamond (maker)                 64.0%  ▁▁▁▁▂▄▅●███▄ beats 127/322  hard · avg 46%          5.0%  sharp · at its level · small run
IFBench (aggregation unspecified)    34.6%  █·█●·██····█ beats 2/5      hard · avg 30%          2.5%  at its level · small run · several runs
MMLU-Pro (maker)                     73.3%  ▁▁·▁▁▁▃▂▄●█▁ beats 95/188   medium · avg 60%        2.3%  at its level · small run
AGIEval English                      83.1%  ▄····▄·▄●··█ beats 3/5      easy · avg 76%          2.1%  typical
ZebraLogic                           60.8%  █▄····●·▄·▄▄ beats 3/6      medium · avg 52%        1.9%  at its level
PopQA                                17.9%  ▄··●···▄···█ beats 1/5      hard · avg 20%          1.6%  at its level
EuroEval HellaSwag (English), tes…   62.4%  ▃·▁▁▂▁▂●█▇▄▁ beats 37/100   medium · avg 67%        0.8%  sharp · at its level · noisy
EuroEval Life in the UK (English …   92.1%  ▁▁▁▁▁▁▁▂▄█●▃ beats 125/146  easy · avg 82%          0.4%  noisy
Humanity's Last Exam (maker)          6.5%  ▁▁▁●▆▆▇██▆▁▃ beats 7/111    very hard · avg 3.8%       —  not counted yet: no other model in the …
## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its *counts* column above; here it is added up by who ran the tests. [Method](../method.md) 
source                                    results  share of the score
Olmo 3                                         14   41% ▇▇▇▇▇
NVIDIA (submitted / maker-reported)            11   38% ▇▇▇▇▇
NVIDIA vendor-reported benchmark scores…        5   13% ▇▇
allenai/Olmo-3-7B-Think model card              2    6% ▇
EuroEval leaderboards                           2    1% ▇
prior: pull toward the middle                        0% ▇
weighted average                                   = 55.4
Licences: Olmo 3 ([CC BY 4.0](../../licenses/report-2512.13961v2.txt)) · NVIDIA ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · allenai/Olmo-3-7B-Think model card ([Apache-2.0](../../licenses/allenai--Olmo-3-7B-Think.txt)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE))
Not counted (7) 
source                  test                            score  why
nvidia/NVIDIA-Nemotro…  AIME 2025                       72.1%  only model in this source's run
nvidia/NVIDIA-Nemotro…  MATH-500                        97.8%  only model in this source's run
nvidia/NVIDIA-Nemotro…  GPQA (subset unspecified)       64.0%  only model in this source's run
nvidia/NVIDIA-Nemotro…  LiveCodeBench (version unspec…  71.1%  only model in this source's run
nvidia/NVIDIA-Nemotro…  BFCL v3                         66.9%  only model in this source's run
nvidia/NVIDIA-Nemotro…  IFEval instruction-level stri…  90.3%  too few models have it yet
nvidia/NVIDIA-Nemotro…  Humanity's Last Exam             6.5%  too few models have it yet
## How to run it ### Through an API api (Amazon Bedrock) · $0.10/M blended · $0.06 in / $0.23 out per 1M tokens · [docs.aws.amazon.com](https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-nvidia-nvidia-nemotron-nano-9b-v2.html) [price source](https://aws.amazon.com/bedrock/pricing): USD per 1M tokens, Amazon Bedrock on-demand 'NVIDIA Nemotron Nano 2' (Bedrock model ID nvidia.nemotron-nano-9b-v2; the separate 'Nano 2 VL' row is the 12B VL model), US East (N. Virginia)/Ohio and US West (Oregon). GovCloud rates excluded. Currency: USD. Amazon Bedrock official pricing page, observed 2026-09-25. · [context source](https://github.com/anomalyco/models.dev/blob/a668908cfe7073faa0363699486fc41a3ff52412/models/nvidia/nemotron-nano-9b-v2.toml): model catalogue context ceiling (not a promise by a hosted API) ### llama.cpp (recommended) 1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md) ```sh # macOS / Linux curl -LsSf https://llama.app/install.sh | sh # Windows irm https://llama.app/install.ps1 | iex ``` 2. Run Nemotron Nano 9B v2 (gguf-q4_k_m, needs ~5.83 GB, from [huggingface.co/bartowski](https://huggingface.co/bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF/tree/main)) ```sh llama cli -hf bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF:Q4_K_M ``` 3. Or start a local server and open the address it prints (http://localhost:9931) ```sh llama serve -hf bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF:Q4_K_M ``` Other versions: gguf-q8_0 (~11.17 GB) `llama cli -hf bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF:Q8_0` ### LM Studio 1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download) 2. Run Nemotron Nano 9B v2 (gguf-q4_k_m, needs ~5.83 GB, from [huggingface.co/bartowski](https://huggingface.co/bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF/tree/main)) ```sh lms get https://huggingface.co/bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF@Q4_K_M -y ``` 3. Then run lms ls to see its name, and lms chat followed by that name Other versions: gguf-q8_0 (~11.17 GB) `lms get https://huggingface.co/bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF@Q8_0 -y` · 6bit (~7.2 GB, Mac only) `lms get https://huggingface.co/mlx-community/NVIDIA-Nemotron-Nano-9B-v2-6bit -y` ### Ollama 1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download) ```sh # Linux curl -fsSL https://ollama.com/install.sh | sh ``` 2. Run Nemotron Nano 9B v2 (gguf-q4_k_m, needs ~5.83 GB, from [huggingface.co/bartowski](https://huggingface.co/bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF/tree/main)) ```sh ollama run hf.co/bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF:Q4_K_M ``` Other versions: gguf-q8_0 (~11.17 GB) `ollama run hf.co/bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF:Q8_0` ### MLX (Mac) 1. Install MLX (Mac). Apple silicon Macs only, text only. ```sh # with uv uv tool install mlx-lm # or pip install mlx-lm ``` 2. Run Nemotron Nano 9B v2 (6bit, needs ~7.2 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/NVIDIA-Nemotron-Nano-9B-v2-6bit/tree/76863058996b1f162fc4db6a0dac595fb7f06954)) ```sh mlx_lm.chat --model mlx-community/NVIDIA-Nemotron-Nano-9B-v2-6bit ``` 3. Or start an OpenAI-compatible server at http://localhost:8080/v1 ```sh mlx_lm.server --model mlx-community/NVIDIA-Nemotron-Nano-9B-v2-6bit ``` Model repository and documentation: [huggingface.co/nvidia](https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2) ## Quantization evidence How much of the full-precision result a quantized version kept, measured within one study. Rows from the same study share a table and aren't independent. They don't describe every download of that precision. 
precision                       test                                full → quantized    kept
NVFP4 PTQ (post-training, W4A…  IFEval (instruction-level) accura…     0.903 → 0.898   99.4%
NVFP4 PTQ (post-training, W4A…  MATH-500 accuracy (NVIDIA evaluat…     0.978 → 0.972   99.4%
NVFP4 PTQ (post-training, W4A…  AIME 2025 accuracy (NVIDIA evalua…     0.711 → 0.698   98.2%
NVFP4 PTQ (post-training, W4A…  GPQA-Diamond accuracy (NVIDIA eva…       0.64 → 0.59   92.2%
NVFP4 QAD (quantization-aware…  AIME 2025 accuracy (NVIDIA evalua…     0.711 → 0.715  100.6%
NVFP4 QAD (quantization-aware…  MATH-500 accuracy (NVIDIA evaluat…     0.978 → 0.972   99.4%
NVFP4 QAD (quantization-aware…  IFEval (instruction-level) accura…     0.903 → 0.893   98.9%
NVFP4 QAD (quantization-aware…  GPQA-Diamond accuracy (NVIDIA eva…      0.64 → 0.627   98.0%
Each kept value links to its source. [interactive version](https://whatmodel.sh/model/nemotron-nano-9b-v2) · [JSON](https://whatmodel.sh/api?format=json&all=1)