TAG
— BROWSING TAG —

#ALIBABA

26
POSTS
AIDeepSeek V4‑Pro 1.6T‑Parameter AI Model Architecture

DeepSeek V4: 1M‑Token Context and Budget Frontier AI Models

DeepSeek V4 launches with 1M‑token context, MoE‑based Pro and Flash models, and prices far below U.S. rivals, focusing on code and agent workloads.

DeepSeekAlibabaOpen Source
AIQwen3.6 Coding Agent Benchmarks Chart Visual

Exploring Qwen3.6: Coding Benchmarks and Speed

Qwen3.6-35B-A3B: Open MoE LLM with 35B params (3B active), tops SWE-bench at 73.4%. Fast local runs at 170 t/s, strong agents vs Gemma4.

QwenAlibabaOpen Source
AIOpenRouter LLM Leaderboard April

Chinese AI Models Dominate OpenRouter Top Six in Token Usage

Chinese LLMs took top 6 on OpenRouter with 13T tokens vs US 3T; Qwen3.6 Plus led at 4.6T. Five weeks straight ahead.

ClaudeAlibabaQwen+3
AIOpenSandbox Logo

OpenSandbox: A Unified Sandbox Layer For AI Agents

Alibaba’s OpenSandbox offers a unified, secure API so AI agents can run code, browse and train in isolated Docker or Kubernetes sandboxes.

AlibabaOpen Source
AIQwen Beats gpt-oss-120B with Laptop Power

Alibaba's Tiny Qwen Beats Big OpenAI Model

Alibaba's Qwen3.5-9B tops OpenAI's gpt-oss-120B on GPQA, Video-MME benchmarks. Runs on laptops, multimodal edge AI star.

AlibabaQwenOpen Source
AIQwen3.5 Medium Model 35B

Qwen3.5 Medium Matches Sonnet 4.5 on Local Hardware

Qwen3.5-27B/35B open-source, runs local on 18-24GB VRAM. Benchmarks rival Claude Sonnet 4.5, 1M ctx, tool-native.

QwenAlibabaOpen Source
AIDeepSeek V4 is Coming

DeepSeek V4 vs Claude/GPT: Strengths, Controversies Ahead

DeepSeek V4 eyes early March drop, topping Claude/GPT in code benchmarks with 1M context, but Blackwell chip scandal brews US ire.

AlibabaDeepSeek
AICinematic AI video frame generated by Seedance 2.0

Seedance 2.0: China’s AI video shocks Hollywood

Seedance 2.0 shows China’s AI video can rival OpenAI and Google, mixing realism with serious ethical and legal questions.

ByteDanceAlibaba
AIQwen3.5‑397B‑A17B model card on ModelScope.

Qwen 3.5: Faster, Cheaper And More Multimodal AI

Alibaba’s Qwen 3.5 boosts multimodal and agent skills, narrowing the gap with US AI leaders while staying efficient through a sparse MoE design.

QwenAlibaba
AI2K Qwen Image 2.0 slide layout with crisp English text

Inside Qwen Image 2.0: Slides, Posters And Photorealism

Qwen Image 2.0 unifies generation and editing with sharper 2K visuals, better humans and typography that finally makes text‑heavy designs workable.

QwenAlibaba
AIQwen3 Max Thinking AI agent visualized with Qwen bear

Qwen3 Max Thinking: New Contender in Deep Reasoning

Alibaba’s reasoning model pairs test‑time scaling with web search and code tools, topping Gemini 3 Pro and GPT‑5.2 on HLE search and math benchmarks.

AlibabaQwen
AIAlibaba Qwen model visual for cross border AI usage.

Why US platforms are turning to Chinese AI models

Chinese open source AI like DeepSeek and Qwen is winning US clients by mixing solid performance with flexibility and lower operating costs.

AlibabaQwenDeepSeek
AIDeepSeek mHC: Smarter Scaling For Giant AI Models

How DeepSeek Trains Powerful Models On A Budget

The mHC idea lets DeepSeek train 3–27B models more efficiently, supporting its push to rival US labs while spending only a fraction on hardware.

AlibabaDeepSeek
AIQwen-Image-2512 vs Nano Banana Pro AI generator visual

Building On Qwen-Image-2512 Instead Of Closed Image AI

Qwen-Image-2512 offers open source, Apache licensed image generation that challenges Google’s Nano Banana Pro for flexible, enterprise ready workflows...

AlibabaQwen
AIZhipu’s GLM 4.7 and MiniMax M2.1 boost coding and agent skills, as Doubao and Yuanbao

Chinese Coding Models GLM 4.7 and M2.1 Step Up

Chinese coding models now sit close to Claude tier performance, while Doubao and Yuanbao prove domestic AI platforms can win huge mainstream audiences...

AnthropicAlibabaDeepSeek+1
AIQwen and DeepSeek vs closed US AI platforms

How Open Source AI Is Quietly Deflating the Hype Today

As enterprises discover capable open models they can run themselves, the pricing power of closed US platforms starts to look increasingly fragile.

AlibabaQwenMeta+1
AIFrom Raster to Layers: What Qwen-Image-Layered Does

Layered AI Images: Inside Qwen-Image-Layered Editing

Qwen-Image-Layered splits flat images into editable RGBA layers, enabling consistent, Photoshop-like edits with precise control over each element.

QwenAlibaba
AIAlibaba’s Z-Image-Turbo excels in photorealistic text image AI

Z-Image-Turbo: Alibaba’s 6B param text-to-image powerhouse

Z-Image-Turbo is a 6B param model by Alibaba for photorealistic, fast text-to-image generation with English and Chinese text support.

Alibaba
AIDeepSeek reveals R1 cost

DeepSeek R1: $294K training cost, ultra‑low inference rates

DeepSeek’s R1 cost ~$294K to train with 512 H800 GPUs and 80h. Inference costs are ~$0.55 input / $2.19 output per million tokens.

DeepSeekAlibaba
AIDeepSeek v.3

DeepSeek-V3 Unveiled: 236B Parameters and 128K Context for Free

DeepSeek-V3 majorly upgrades from V2 with a massive 236B parameter count, a vastly larger 128K context window, and superior performance, all offered f...

AlibabaDeepSeek
20 / 26