2026-09-19 04:51
AI123

Cognition发布SWE-2模型,称多项基准测试比肩前沿、成本最高降低70%

Cognition宣布推出SWE-2模型,称这是其至今最接近前沿水平的模型。在多项基准测试中,SWE-2与近期前沿模型表现持平,成本最高可降低70%。

Cognition表示,团队将强化学习扩展到数万亿参数的规模,并通过调整训练策略,在能力和成本两方面推动帕累托曲线优化。

Cognition团队成员Silas Alberti在社交平台透露,SWE-2在桌面端和命令行环境中的需求空前,公司已为此获得更多算力资源。为满足用户需求,即日起至2026年10月8日,Devin Cloud中的SWE-2将对Pro、Max及Teams订阅用户免费开放。

来源
  1. 2026-09-19 04:16 | X:RT Silas Alberti: Demand for SWE-2 in Desktop & CLI is unprecedented. Glad to announce, we were able to secure additional compute. We're making SWE-2 ...
    阅读原文

    RT Silas Alberti Demand for SWE-2 in Desktop & CLI is unprecedented. Glad to announce, we were able to secure additional compute. We're making SWE-2 free in Devin Cloud for Pro, Max & Teams subscribers - until October 8. Enjoy! Cognition: Introducing SWE-2, our closest model yet to the frontier. On ...

其他新闻

查看全部