open-source AI Articles | Developments Today

$Alibaba's open Qwen 3.5 takes aim at GPT-5 mini and Claude Sonnet 4.5 at a fraction of the cost$

Alibaba تطلق نماذج Qwen 3.5 المفتوحة للتحدي ضد GPT-5 Mini و Claude Sonnet 4.5

قدمت Alibaba سلسلتها الجديدة من نماذج Qwen 3.5، وهي عائلة من أربعة نماذج ذكاء اصطناعي مفتوحة تقول الشركة إنها تنافس GPT-5 mini و Claude Sonnet 4.5 بتكلفة أقل بكثير. تشمل المجموعة نموذج Flash خفيف الوزن وثلاثة نماذج مزيج الخبراء (mixture-of-experts) التي تغطي مقاييس معلمات مختلفة.

Key Takeaways

أطلقت Alibaba أربعة نماذج Qwen 3.5، بما في ذلك متغيرات Flash و 35B-A3B و 122B-A10B و 27B
تستخدم النماذج بنية مزيج الخبراء لتقديم أداء عالٍ بتكلفة حوسبة أقل
تدعي Alibaba أداءً تنافسيًا مع GPT-5 mini و Claude Sonnet 4.5
تم إصدار جميع النماذج كأوزان مفتوحة، مما يتحدى نموذج المصدر المغلق

DE

DT Editorial AI·Feb 26, 2026·via the-decoder.com

Abstract visualization of neural network connections and artificial intelligence

AI & Robotics

NousCoder-14B: Open-Source Coding Model Takes On Claude Code

Nous Research releases NousCoder-14B, an open-source coding model that benchmarks competitively against proprietary tools like Claude Code.

Key Takeaways

Nous Research releases NousCoder-14B, achieving 42.8% on SWE-bench Verified versus Claude Code's 54% at a fraction of the size
The 14B parameter model runs on consumer GPUs with 24GB VRAM under Apache 2.0 license for unrestricted commercial use
The narrowing gap between open-source and proprietary coding models raises questions about the future of paid AI coding tools

DE

DT Editorial AI·Feb 11, 2026·via nousresearch.com

Nous Research's NousCoder-14B is an open-source coding model landing right in the Claude Code moment

AI & Robotics

NousCoder-14B Brings Fully Open-Source RL to Competitive Coding

Nous Research releases a 14-billion-parameter coding model trained with reinforcement learning in just four days, publishing the entire training stack.

Key Takeaways

NousCoder-14B achieves 67.87% on LiveCodeBench v6, a 7-point improvement over its Qwen3-14B base model through reinforcement learning
The entire training pipeline, RL environment, and benchmark suite are published for full reproducibility by any researcher
Training took just four days on 48 Nvidia B200 GPUs using Nous Research's Atropos reinforcement learning framework

DE

DT Editorial AI·Feb 11, 2026·via venturebeat.com

#open-source AI

Alibaba تطلق نماذج Qwen 3.5 المفتوحة للتحدي ضد GPT-5 Mini و Claude Sonnet 4.5

NousCoder-14B: Open-Source Coding Model Takes On Claude Code

NousCoder-14B Brings Fully Open-Source RL to Competitive Coding

Alibaba تطلق نماذج Qwen 3.5 المفتوحة للتحدي ضد GPT-5 Mini و Claude Sonnet 4.5

NousCoder-14B: Open-Source Coding Model Takes On Claude Code

NousCoder-14B Brings Fully Open-Source RL to Competitive Coding