DeepSeek, a China-based AI startup, has released two new reasoning-first models, V3.2 and V3.2-Special, that the company says rival the latest offerings from OpenAI and Google. The announcement ...

The files in this repository extracted from CodeForces using codeforces2pdf, folders names represent the contests` IDs from CodeForces URLs, each folder contains the problems of the contest.

OpenAI and NVIDIA have unveiled two cutting-edge open-weight large language models (LLMs) — gpt-oss-120b and gpt-oss-20b — designed to bring advanced reasoning capabilities into the hands of ...

📌 A curated collection of daily Codeforces problem solutions implemented in C++, focusing on clean code, optimized approaches, and competitive programming techniques. 📖 About This repository is ...

The Test of Mathematics for University Admission (TMUA) tests your mathematical thinking and reasoning skills. Which LSE programmes use the TMUA in their selection process? For these two programmes ...

为了解决效率瓶颈，小红书大模型基建团队提出新架构 PIPO（Pair-In, Pair-Out）：在模型输入侧，压缩器把两个输入 token 折叠成一个 latent；在模型输出侧，MTP head 把一个隐藏状态展开成一个额外输出 token；进而实现减半输入长度、双倍每步输出长度。 长链路推理已 ...

We are home to 30 faculty, four staff, approximately 60 graduate students, several research visitors, and numerous undergraduate students. We offer exciting and challenging programs leading to BMath, ...

这项由谷歌DeepMind研究团队主导开发的技术报告于2026年7月2日正式发布，论文编号为arXiv:2607.02770v1，有兴趣深入探究的读者可通过该编号在arXiv平台检索完整技术文档。