发生了什么: Introducing SWE-2 Flash

Cognition released SWE-2 Flash, a smaller coding model achieving 20% on Terminal-Bench 4.0, up from 0.9% baseline, at one-fifth the cost.

为何重要: Introducing SWE-2 Flash

Engineers building software agents, automations, and code review tools who need faster inference and lower costs than larger models.

注意: Introducing SWE-2 Flash

Weights are not publicly available; access requires approval from Cognition's partner program due to cybersecurity concerns and potential misuse risks.

新闻

Introducing SWE-2 Flash

Cognition · 发布于 2026年11月4日 · 阅读约3分钟

30秒读懂

发生了什么: Cognition released SWE-2 Flash, a smaller coding model achieving 20% on Terminal-Bench 4.0, up from 0.9% baseline, at one-fifth the cost.
为何重要: Engineers building software agents, automations, and code review tools who need faster inference and lower costs than larger models.
注意: Weights are not publicly available; access requires approval from Cognition's partner program due to cybersecurity concerns and potential misuse risks.

阅读完整文章

这条新闻背后的模式

SWE-bench Suite
Terminal-Bench
SWE-bench Pro

每个模式都讲清楚技术如何运作、何时值得投入，以及在哪里会失效。

The Agent Architect

每周一个模式、一个权衡、一个生产事故案例。为构建智能体系统的人准备的每周简报。

每周一封邮件,一键退订。您的地址仅用于发送简报。