Sky-T1-32B-Preview

Home/Models/NovaSky-AI/Sky-T1-32B-Preview

NovaSky-AI/Sky-T1-32B-Preview

kind: original
quant: bf16 (16-bit)
family: misc
modality: text-generation
root model: Qwen/Qwen2.5-32B-Instruct
provider: NovaSky-AI
license: apache-2.0
language: en
context: —
library: transformers
vRAM (GB): —
size on disk: —
downloads: 31
likes: 550
created: —
last modified: 2025-01-13T08:10:51+00:00 (2025)
huggingface: https://huggingface.co/NovaSky-AI/Sky-T1-32B-Preview
tags: transformerssafetensorsqwen2text-generationconversationalendataset:codeparrot/appsdataset:BAAI/TACOdataset:AI-MO/NuminaMath-CoTarxiv:2412.09413base_model:Qwen/Qwen2.5-32B-Instructbase_model:finetune:Qwen/Qwen2.5-32B-Instructlicense:apache-2.0text-generation-inferenceendpoints_compatibleregion:usdeploy:azure

benchmark	category	score	context	conf	evidence
LiveCodeBench	coding	86.3	LiveCodeBench-Easy	medium	`\| LiveCodeBench-Easy \| 86.3 \| 84.6 \| 90.7 \| 92.9 \|`
MATH-500	math	82.4	Math500	medium	`\| Math500 \| 82.4 \| 76.2 \| 85.4 \| 81.4 \|`
LiveCodeBench	coding	56.8	LiveCodeBench-Medium	medium	`\| LiveCodeBench-Medium \| 56.8 \| 40.8 \| 56.3 \| 54.9 \|`
GPQA Diamond	reasoning	56.8	GPQA-Diamond	medium	`\| GPQA-Diamond \| 56.8 \| 45.5 \| 52.5 \| 75.2 \|`
AIME 2024	math	43.3	AIME2024	medium	`\| AIME2024 \| 43.3 \| 16.7 \| 50.0 \| 40.0 \|`
LiveCodeBench	coding	17.9	LiveCodeBench-Hard	medium	`\| LiveCodeBench-Hard \| 17.9 \| 9.8 \| 17.1 \| 16.3 \|`

none scraped.