NVIDIA-Nemotron-Nano-9B-v2-Japanese

Home/Models/nvidia/NVIDIA-Nemotron-Nano-9B-v2-Japanese

nvidia/NVIDIA-Nemotron-Nano-9B-v2-Japanese

kind: original
quant: bf16 (16-bit)
family: nemotron
modality: text-generation
root model: nvidia/NVIDIA-Nemotron-Nano-9B-v2
provider: nvidia
license: other
language: en, ja
context: 128K
library: transformers
vRAM (GB): —
size on disk: —
downloads: 4,431
likes: 140
created: —
last modified: 2026-04-21T05:39:44+00:00 (2026)
huggingface: https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2-Japanese
tags: transformerssafetensorsnemotron_htext-generationnvidiapytorchconversationalcustom_codeenjadataset:nvidia/Nemotron-Personas-Japandataset:nvidia/Nemotron-CC-v2.1dataset:nvidia/Nemotron-Pretraining-Specialized-v1dataset:nvidia/Nemotron-Agentic-v1dataset:nvidia/Nemotron-Instruction-Following-Chat-v1dataset:globis-university/aozorabunko-cleandataset:HuggingFaceFW/fineweb-2arxiv:2504.03624arxiv:2508.14444base_model:nvidia/NVIDIA-Nemotron-Nano-9B-v2base_model:finetune:nvidia/NVIDIA-Nemotron-Nano-9B-v2license:otherendpoints_compatibleregion:us

benchmark	category	score	context	conf	evidence
MT-Bench	chat	89.2	MT-Bench	medium	`\| MT-Bench \| 0.892 \| 0.906 \|`
BFCL (function calling)	agentic	64.9	BFCL v3	medium	`\| BFCL v3 \| 0.649 \| 0.608 \|`
IFEval	instruction-following	63.2	M-IFEval	medium	`\| M-IFEval \| 0.632 \| 0.619 \|`
TruthfulQA	safety	49.8	JtruthfulQA	medium	`\| JtruthfulQA \| 0.498 \| 0.433 \|`
ARC-AGI	reasoning	6.0	ARC-AGI 1/2	medium	`\| ARC-AGI 1/2 \| 0.060 \| 0.070 \|`
Humanity's Last Exam	knowledge	5.7	HLE	medium	`\| HLE \| 0.057 \| 0.026 \|`

none scraped.