英伟达发布财报之前 DeepSeek版本升级

Core Insights - DeepSeek has announced the completion of a minor version upgrade for its R1 model, inviting users to test the new features on its official platforms [1][3] - The company, based in Hangzhou, gained significant attention in January when its original R1 model outperformed Western counterparts in standardized evaluations, with development costs reportedly in the millions [3] - The recent upgrade was announced just hours before NVIDIA's latest earnings report, which had previously seen its stock price decline due to the R1 model's initial release [3] Upgrade Highlights - The upgraded R1 model now performs deep reasoning similar to Google models [4][7] - Improvements in writing tasks have made outputs more natural and better formatted [4][7] - The model exhibits a distinct reasoning style that is both fast and thoughtful [4][7] - Users can engage in long thinking sessions, with each task taking up to 30-60 minutes [4][7]