2025年2月11日 · Following, we show our evaluation results for the Sky-T1-32B-Preview model across math, coding, and science benchmarks. We also evaluate on non-reasoning benchmarks (these are benchmarks for instruction-following, QA, etc) to test whether the model has traded-off capability in other domains for better performance in reasoning-related benchmarks.

华诺星空技术股份有限公司成立于2006年，是我国实体空间安全领域的领军企业。 公司聚焦“智能安防+AI”，现已形成以全域安防、安全应急、智能感知、智能系统、智慧交通等五大产品线为主的产品族群。 源源不断地为应急、公安、司法、自然资源、水利、国动、交通、电力、石化等行业的客户提供领先的产品和解决方案。 您可以点击下方按钮，进行需求评估。 Copyright©2022 华诺星空技术股份有限公司 All Rights Reserved. 备案号 湘ICP备12001712号. 华诺星空技术股份有 …

Novasky Technology Company Limited by Shares was established in 2006, is the leader in the field of physical space security. Novasky focuses on "intelligent security + AI", and has now formed a product group with four major product lines, including Safety Emergency, All-Terrain Security, Drone Applications and IntelliSense.

Sky-T1是加州大学伯克利分校Sky Computing实验室的研究团队NovaSky发布的开源推理AI模型，名为Sky-T1-32B-Preview。 是首个开源推理模型，训练数据集和代码均已公开，用户可以从零开始复现该模型。

2025年1月13日 · NovaSky (Next-generation Open Vision and AI) is a collaborative initiative led by students and advisors at UC Berkeley’s Sky Computing Lab. We aim to build the next-generation AI models that are SoTA, open-source and affordable for everyone.

2025年1月10日 · Introducing S*: Test-Time Scaling for Code Generation, start of our releases in the coding domain. S* enables (1) non-reasoning models surpass reasoni...

Developed by: NovaSky Team from Sky Computing Lab at UC Berkeley. 17K verified correct responses from Qwen/QwQ-32B-Preview on coding, math. In addition, we add the science portion from the Still-2 paper. We perform supervised fine tuning on the data, with a batch size of 96. We use Llama-Factory for training.

NovaSky has 2 repositories available. Follow their code on GitHub.

Novasky Technology Company Limited by Shares was established in 2006. The company focuses on electromagnetic detection, sensing and control technology, and provides product and system integration services for anti-terrorism, search&rescue, high-end security and automotive safety.

Our model training was completed in 19 hours on 8 H100 GPUs using DeepSpeed Zero-3 offloading, costing approximately $450 as per Lambda Cloud pricing. Following, we show our …