DeepSeek-V4.1-Flash-1017 released, with simultaneous improvements in benchmark performance and cost
DeepSeek has released DeepSeek-V4.1-Flash-1017, claiming stronger benchmark performance, faster task completion, and shorter outputs, achieving a Pareto improvement in performance and cost; it is now live on the official website, App, and API. This version has been optimized for DeepSeek Harness, Tencent CodeBuddy, WorkBuddy, and others, and requires upgrading to DeepSeek Harness v0.2.1-rc.1 or later for the best experience. Nathan Lambert thinks its naming approach (a minor version number plus a date tag) is quite absurd, but the benchmark gains are large, and with built-in vision capability it may be stronger than GLM 5.3.
Why read this
SelectedAI score 67
The new DeepSeek version improves on both benchmarks and cost at the same time, and the author offers a judgment on its naming approach and built-in vision capability.
Topics
View original post🚀 DeepSeek-V4.1-Flash-1017 is here! Stronger benchmark performance, faster task completion, and shorter outputs than the previous version—a Pareto improvement in performance and cost. Now live on our website, app, and API. 🛠️ Optimized for @DeepSeekHarness, Tencent CodeBuddy, WorkBuddy, and more. Upgrade to DeepSeek Harness v0.2.1-rc.1 or later for the best experience with the new model. 1/4
View original
🚀 DeepSeek-V4.1-Flash-1017 is here! Stronger benchmark performance, faster task completion, and shorter outputs than the previous version—a Pareto improvement in performance and cost. Now live on our website, app, and API. 🛠️ Optimized for @DeepSeekHarness, Tencent CodeBuddy, WorkBuddy, and more. Upgrade to DeepSeek Harness v0.2.1-rc.1 or later for the best experience with the new model. 1/4
Body
This naming scheme is bonkers (minor model numbers and date tags) and the benchmark step is big. Could prove stronger than GLM 5.3 with vision built in.
The King of open weights is back?