Composer 2 正式发布

一项针对500家公司开发者使用Cursor的八个月研究发现，在Opus 4.5和GPT-5.2等先进模型发布后，人均周AI使用量增长44%。开发者初期用更强模型完成更多同复杂度任务，4-6周后开始转向更高复杂度工作，高复杂度任务量激增68%，远超低复杂度任务的22%。媒体广告、软件工具和金融科技行业增长最为显著。任务分布呈现结构性变化：文档编写、架构设计等管理性任务增长超50%，而UI设计等独立任务仅增15%，表明开发者角色正从代码生成转向代码库管理。研究揭示了类似杰文斯悖…

针对Blackwell GPU上的小批量解码，研究提出了一种名为"warp decode"的新方法。该方法颠覆了传统以专家为中心的计算路径，改为让每个GPU warp负责计算一个输出神经元。这一根本性改变消除了原有流程中五个纯数据管理的"簿记"步骤，将整个MoE计算层压缩为仅两个内核。其优势在于避免了填充、分散和中间缓冲区的读写，并通过warp独立性实现了更好的调度。在Blackwell GPU上，该方法实现了1.84倍的吞吐量提升，同时输出精度更高，与全FP32参考值的差…

本报告介绍了代码模型Composer 2的训练过程。该模型基于开源基础模型Kimi K2.5，通过两阶段训练：首先进行侧重代码的持续预训练以深化编码知识，随后在高度模拟真实Cursor环境的大规模强化学习中提升端到端智能体性能。在自建的真实任务评估集CursorBench上，Composer 2得分为61.3，较前代提升37%，与前沿模型性能相当。在公开基准SWE-bench Multilingual和Terminal-Bench上分别获得73.7和61.7分，并在保持高精…

在无中央协调器的开放世界多智能体环境Station中，来自不同模型家族的AI智能体自主选择研究方向、开展实验并构建共享科学文献。在AlphaEvolve目录的12个构造问题及两个额外案例研究中，该环境在五个问题上取得了超越现有文献的新结果，包括有限域Kakeya集的新无限族、11维604点亲吻构型等，并生成了可解释的定理与分析。所有原始智能体对话、证明和验证代码均已公开。

斯坦福大学研究人员领衔发布 Terminal-Bench-Science 0.1，用来自生命、物理、地球、数学和工程科学的 70 个专家精选任务评估 AI 智能体的科研能力。

Recent advances in inference-time scaling have significantly improved the reasoning performance of large language models (LLMs). However, these methods typically rely on repeated generation or external verification. To address this limitat…

Agent skills package specialized knowledge and workflows into reusable resources that extend AI agent capabilities. Recent work automatically discovers such skills from agent experience, which enables agents to progressively adapt through …

Recent prominent post-training methods, such as Reinforcement Learning (RL) and On-Policy Self-Distillation (OPSD), have driven rapid progress in mathematical reasoning for large language models, yet their reliance on ground-truth labels p…

To improve large language models' ability to resolve real-world software issues, prior work has focused on constructing large-scale agent trajectory datasets and performing supervised fine-tuning (SFT) on successful trajectories. However, …

In real-world software development, code review typically involves iterative interactions between developers and reviewers to improve software quality, making the process costly and time-consuming. Although recent work explores large langu…