Cognition releases early preview of SWE-1.6 and shares RL training research progress
Original title:An Early Preview of SWE-1.6 and Research Update
AI summary
Cognition released an early preview of SWE-1.6. The current checkpoint scores 51.7% on SWE-Bench Pro, 11% higher than SWE-1.5, at the same 950 tok/s speed, and is available to a small group of users for early access.
Key Takeaways
- Cognition released an early preview of SWE-1.
- The current checkpoint scores 51.
- 7% on SWE-Bench Pro,11% higher than SWE-1.
- 5,at the same 950 tok/s speed,and is available to a small group of users for early access.
Quick Facts
- Published
- Updated
- Source
- Cognition 模型 / Devin 博客(网页)
- Category
- Model Releases
- AI score
- 60
官方给出 SWE-1.6 预览版成绩与 RL 训练细节,并提出基准之外的 Model UX 问题,对做智能体训练的人有参考价值。
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Cognition releases early preview of SWE-1.6 and shares RL training research progress. March 1, 2026. https://new.nat6.net/en/news/cmtyo2ihn03d5rog0lfpkae4o/