چین
@arena
بچههای دیپسیک دوباره گل کاشتن و مدل جدیدشون یعنی DeepSeek-V4.1-Flash رسماً رتبه اول مدلهای متنباز رو توی بنچمارک توسعه وب کد آرنا مال خودش کرد




بچههای دیپسیک دوباره گل کاشتن و مدل جدیدشون یعنی DeepSeek-V4.1-Flash رسماً رتبه اول مدلهای متنباز رو توی بنچمارک توسعه وب کد آرنا مال خودش کرد. این شاهکار تازه نفس با ۱۶۷۰ امتیاز، نهتنها رقیب سرسختش کیمی K3 رو جا گذاشته، بلکه جلوتر از کلود اوپوس ۵ ایستاده و فاصلهش با نسخه تاپ GPT-6 هم فقط در حد ۱۰ امتیازه. برگ برنده اصلیش هم اینجاست که با هزینه ده برابر کمتر و فقط با ۳۰ سنت تا ۱.۲ دلار برای هر میلیون توکن، هم خروجی سریعتر و تر و تمیزتری تحویل میده و هم هوای جیب برنامهنویسها رو حسابی داره.
منبع
Arena.ai@arena
Exciting news: DeepSeek-V4.1-Flash-20261017 (Max) by @deepseek_ai is now the #1 open model in Code Arena: WebDev, surpassing Kimi K3 (Max) at 10x lower cost!
Highlights:
- #1 open model, ahead of Kimi K3 (Max)
- #12 overall with 1670 pts, ahead of Claude Opus 5 (High) and within ~10 pts of GPT-6 Sol (Max)
- Shifts the Pareto frontier at $0.3/$1.2 per million input/output tokens
- +52 pts vs. the previous DeepSeek-V4.1-Flash-20260910
Congrats to the @deepseek_ai team on another strong contribution to the open source ecosystem!
DeepSeek @deepseek_ai
🚀 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
۲۶ مهر ۱۴۰۵مشاهده در X


نظرها
هنوز نظری نیست.
بخش نظرها فعلاً در دسترس نیست.