NFT-7B

Home/Models/nvidia/NFT-7B

nvidia/NFT-7B

kind: original
quant: bf16 (16-bit)
family: misc
modality: text-generation
root model: Qwen/Qwen2.5-Math-7B
provider: nvidia
license: other
language: en
context: 32,768
library: transformers
vRAM (GB): —
size on disk: —
downloads: 67
likes: 3
created: —
last modified: 2025-07-15T09:31:57+00:00 (2025)
huggingface: https://huggingface.co/nvidia/NFT-7B
tags: transformerssafetensorsqwen2text-generationnvidiamathreasoningpost-trainingqwenconversationalendataset:BytedTsinghua-SIA/DAPO-Math-17karxiv:2505.18116base_model:Qwen/Qwen2.5-Math-7Bbase_model:finetune:Qwen/Qwen2.5-Math-7Blicense:othertext-generation-inferenceendpoints_compatibleregion:us

benchmark	category	score	context	conf	evidence
MATH-500	math	83.2	MATH500	medium	`\| MATH500 \| 83.2% \| 69.0% \| +14.2% \|`
Math Olympiad	math	47.3	OlympiadBench	medium	`\| OlympiadBench \| 47.3% \| 34.7% \| +12.6% \|`
MATH	math	40.8	Minerva Math	medium	`\| Minerva Math \| 40.8% \| 21.3% \| +19.5% \|`
AIME 2024	math	32.0	AIME24 (avg@32)	medium	`\| AIME24 (avg@32) \| 32.0% \| 13.3% \| +18.7% \|`
AIME 2025	math	18.3	AIME25 (avg@32)	medium	`\| AIME25 (avg@32) \| 18.3% \| 5.5% \| +12.8% \|`
MATH	math	13.3		low	`\| AIME24 (avg@32) \| 32.0% \| 13.3% \| +18.7% \|`

none scraped.