Knowledge catalogue
Cognition AI has published a technical report detailing the design methodology behind their reward system, likely related to their Devin AI software engineering agent. The report covers the performanc
The second technique was two-phase post-training. We first trained purely for capability, then added a latency penalty calibrated from real dogfooding data based on the CDF of how long users stay on S
Today we're releasing SWE-check, a specialized bug detection model we RL-trained with @appliedcompute that matches frontier performance on internal in-distribution evals and makes meaningful progress
GLM-5.1, developed by Z.ai (formerly Zhipu AI), is Z.ai's next-generation flagship model for agentic engineering, featuring 'significantly stronger coding capabilities than its predecessor,' bette...
92% of COBOL developers are expected to retire in the next four years, and 68% of enterprise COBOL modernizations are failing. Agents can help – we wrote up how: https://cognition.ai/blog/how-devin-is
Cognition is live in Japan! 🇯🇵 We launched today with Tokyo Merge - a standing room only event with industry leaders across the country. We’re hiring, partnering, and building in Japan. If you’re buil
Cognition's AI software engineer, Devin, was tasked at 1:40 AM by an OpenSea team member to build the first NFT mint tab for the OpenSea mobile app. Within 24 hours of receiving the request, Devin...
SWE-1.6 is lightning fast! Here's what 950 tok/s feels like - available in Windsurf today. Media We’re releasing SWE-1.6, our best model in both intelligence & model UX. SWE-1.6 matches our Preview mo
A user on X (@willebrew) shared positive impressions of Cognition's SWE-1.6 Fast model, noting that a prototype UI it generated surpassed the output of Figma Make. SWE-1.6 is Cognition's latest mo...
Cognition released SWE-1.6, their latest software engineering model optimized for both intelligence and 'model UX,' now generally available in the Windsurf IDE. It is free for the next three month...
SWE-1.6 is free for everyone in Windsurf for the next 3 months at 200 tok/s. For paying users, we've partnered with Cerebras to serve the model at 950 tok/s. More technical details about this training
We’re releasing SWE-1.6, our best model in both intelligence & model UX. SWE-1.6 matches our Preview model on SWE-Bench Pro while dramatically improving on various behavioral axes. It’s available toda
With SWE-1.6 we've made significant progress on 'intelligence per token'. We post-trained the model from scratch (same pre-trained model) with a similar recipe as SWE-1.6 Preview. Our latest algorithm