DeepSeek-V3-0324-NVFP4

Home/Models/nvidia/DeepSeek-V3-0324-NVFP4

nvidia/DeepSeek-V3-0324-NVFP4

kind: nvfp4
quant: nvfp4 (4-bit)
family: deepseek
modality: text-generation
root model: deepseek-ai/DeepSeek-V3-0324
provider: nvidia
license: mit
language: —
context: 128K
library: Model Optimizer
vRAM (GB): —
size on disk: —
downloads: 31,794
likes: 17
created: —
last modified: 2025-08-22T18:03:46+00:00 (2025)
huggingface: https://huggingface.co/nvidia/DeepSeek-V3-0324-NVFP4
tags: Model Optimizersafetensorsdeepseek_v3nvidiaModelOptDeepSeekV3quantizedFP4text-generationconversationalcustom_codebase_model:deepseek-ai/DeepSeek-V3-0324base_model:finetune:deepseek-ai/DeepSeek-V3-0324license:mit8-bitregion:us

benchmark	category	score	context	conf	evidence
MATH-500	math	94.0	MATH-500	medium	`\| MATH-500 \| 94 \| 94.4 \|`
MGSM	math	92.0	MGSM	medium	`\| MGSM \| 92 \| 92.8 \|`
MMMU	multimodal	82.0	MMMU Pro	medium	`\| MMMU Pro \| 82 \| 82.9 \|`
GPQA Diamond	reasoning	66.0	GPQA Diamond	medium	`\| GPQA Diamond \| 66 \| 67.2 \|`
AIME 2024	math	52.0	AIME 2024	medium	`\| AIME 2024 \| 52 \| 49.3 \|`
LiveCodeBench	coding	41.0	LiveCodeBench	medium	`\| LiveCodeBench \| 41 \| 52.23 \|`

none scraped.