Title: 悟空率先接入国产最强编程模型 Qwen3.6-Plus | BestBlogs.dev
URL Source: https://www.bestblogs.dev/article/f58d8245
Published Time: 2026-04-02 08:28:38
Markdown Content: Skip to main content Toggle navigation menu Toggle navigation menuArticlesPodcastsVideosTweetsSourcesNewsletters
⌘K
Change language Switch ThemeSign In
Narrow Mode
悟空率先接入国产最强编程模型 Qwen3.6-Plus
量 量子位 @梦晨
One Sentence Summary
Alibaba has released the Qwen3.6-Plus model, which has been integrated into its enterprise AI application 'Wukong,' significantly enhancing agentic programming, long-horizon task planning, and cost-effectiveness.
Summary
Alibaba has officially launched its new-generation large language model, Qwen3.6-Plus, with its flagship enterprise AI application 'Wukong' already integrated. The model performs exceptionally well in authoritative programming benchmarks like SWE-bench, surpassing domestic models such as GLM-5 and Kimi-K2.5, and approaching the level of the Claude series. With the new model, 'Wukong' has evolved from an AI 'copilot' to a 'collaborator' capable of handling tasks independently, supporting autonomous cross-file code writing and iterative debugging. Furthermore, the model enhances long-horizon task planning, making it suitable for complex enterprise scenarios such as R&D, legal, and finance. Additionally, the input cost has been drastically reduced to 2 RMB per million tokens, significantly improving the cost-effectiveness for large-scale enterprise applications.
Main Points
* 1. Qwen3.6-Plus model released, marking a significant breakthrough in domestic programming and agent capabilities.It performs exceptionally well in benchmarks like SWE-bench and Claw-Eval, with programming capabilities that surpass similar domestic models and rival the Claude series, while strengthening native multimodal and reasoning performance. * 2. Alibaba's 'Wukong' achieves the transition from an AI assistant to an independent 'collaborator'.Leveraging the new model, Wukong can autonomously complete the entire process from solution breakdown and cross-file code writing to testing and verification, significantly enhancing its ability to handle complex, long-horizon enterprise tasks. * 3. Combining high cost-effectiveness with scenario understanding to accelerate the adoption of enterprise Agents.With input costs as low as 2 RMB per million tokens, combined with the domain knowledge accumulated by DingTalk in the enterprise market, AI capabilities are precisely embedded into real-world workflows such as R&D, legal, and finance.
Metadata
AI Score
86
Website qbitai.com
Published At Today
Length 871 words (about 4 min)
Sign in to use highlight and note-taking features for a better reading experience. Sign in now
4月2日，阿里巴巴正式发布新一代大语言模型Qwen3.6-Plus，阿里在企业级市场的旗舰AI应用悟空率先完成接入。
Qwen3.6-Plus在代码、智能体、推理、原生多模态等能力上整体性能大幅增强，在智能体编程SWE-bench系列评测、真实世界智能体任务Claw-Eval等权威评测中，千问3.6编程表现超越2倍乃至3倍参数量的GLM-5、Kimi-K2.5等模型，成为当下编程能力最强的国产模型，接近全球最强编程模型 Claude 系列。
对悟空用户而言，使用新模型可带来三大提升：智能体编程能力对比其他通用模型大幅提升，AI从”副驾驶”进阶为能独立承担子任务的”协作者”，可自主编写跨文件代码、运行测试并迭代修复；长程任务规划与执行能力显著增强，面对企业场景中普遍的多步骤复杂工作流，能更可靠地拆解目标、调度资源并交付成果；同时性价比优势突出，每百万 Tokens 输入低至2元，大幅降低企业规模化使用门槛。
悟空作为阿里巴巴企业级 AI旗舰应用，承担着将模型最新技术成果与企业真实需求对接的关键角色，已率先接入 Qwen3.6-Plus，为用户带来编程能力更强、执行效果更优、性价比更高的企业级 Agent 使用体验。
例如在一人开发等研发类场景，用户可用自然语言描述需求，悟空自主完成从方案拆解、代码编写、网站生成到测试验证的全流程；在一人法务等知识管理类场景，悟空能跨文档提取关键信息、对比条款变化、识别风险并生成决策建议；在财务、制造等业务流程自动化场景，员工只需提出目标，悟空即可自主完成跨系统的查询、填报、审批等操作，将繁琐流程变为“一句话”的智能服务。
悟空率先接入Qwen3.6-Plus，是阿里AI模型与场景结合的关键实践。
阿里悟空产品负责人表示，企业级AI应用的竞争，本质上是”模型能力×场景理解”的乘积。Qwen3.6-Plus在智能体编程和长程任务规划的突破，为悟空解决企业客户复杂痛点提供了新的基础，而钉钉多年深耕企业市场所积累的场景认知，让悟空知道如何将这些能力精准嵌入真实工作流。
即日起，悟空会员用户在 v0.9.20 及更高版本中，可抢先使用Qwen3.6-Plus 模型。
量 量子位 @梦晨
One Sentence Summary
Alibaba has released the Qwen3.6-Plus model, which has been integrated into its enterprise AI application 'Wukong,' significantly enhancing agentic programming, long-horizon task planning, and cost-effectiveness.
Summary
Alibaba has officially launched its new-generation large language model, Qwen3.6-Plus, with its flagship enterprise AI application 'Wukong' already integrated. The model performs exceptionally well in authoritative programming benchmarks like SWE-bench, surpassing domestic models such as GLM-5 and Kimi-K2.5, and approaching the level of the Claude series. With the new model, 'Wukong' has evolved from an AI 'copilot' to a 'collaborator' capable of handling tasks independently, supporting autonomous cross-file code writing and iterative debugging. Furthermore, the model enhances long-horizon task planning, making it suitable for complex enterprise scenarios such as R&D, legal, and finance. Additionally, the input cost has been drastically reduced to 2 RMB per million tokens, significantly improving the cost-effectiveness for large-scale enterprise applications.
Main Points
* 1. Qwen3.6-Plus model released, marking a significant breakthrough in domestic programming and agent capabilities.
It performs exceptionally well in benchmarks like SWE-bench and Claw-Eval, with programming capabilities that surpass similar domestic models and rival the Claude series, while strengthening native multimodal and reasoning performance.
* 2. Alibaba's 'Wukong' achieves the transition from an AI assistant to an independent 'collaborator'.
Leveraging the new model, Wukong can autonomously complete the entire process from solution breakdown and cross-file code writing to testing and verification, significantly enhancing its ability to handle complex, long-horizon enterprise tasks.
* 3. Combining high cost-effectiveness with scenario understanding to accelerate the adoption of enterprise Agents.
With input costs as low as 2 RMB per million tokens, combined with the domain knowledge accumulated by DingTalk in the enterprise market, AI capabilities are precisely embedded into real-world workflows such as R&D, legal, and finance.
Key Quotes
* AI has evolved from a 'copilot' to a 'collaborator' capable of independently handling subtasks, autonomously writing cross-file code, running tests, and performing iterative repairs. * The competition in enterprise AI applications is essentially the product of 'model capability × scenario understanding'. * The breakthroughs of Qwen3.6-Plus in agentic programming and long-horizon task planning provide a new foundation for Wukong to solve complex pain points for enterprise clients.
AI Score
86
Website qbitai.com
Published At Today
Length 871 words (about 4 min)
Tags
Qwen3.6-Plus
Alibaba
Wukong
Agentic Programming
Large Language Model
Related Articles
* Full Transcript | DingTalk CEO Wu Zhao: Breaking Down DingTalk, Rebuilding with AI, and Forging 'Wukong' * Yao Shunyu Lectures Face-to-Face with Tang Jie, Yang Zhilin, and Lin Junyang! Four Schema Heroes Debate Heroes at Zhongguancun * AI Starts to "Take Action", Alibaba's Qwen Leads the World * Zhipu GLM-5 Technical Deep Dive: Full Compatibility with Domestic Chips Like Huawei Sparks Global Discussion and asynchronous reinforcement learning infrastructure, while completing full-stack adaptation for domestic chips.") * Google's New Paper Crashes Memory Stocks! 6x KV Cache Compression, Netizens: Silicon Valley Becomes Reality * Solving Issues Overnight: Practice of 'Diagnosis' Agents in the E-commerce Transaction Domain * GPT-5.4 Released: OpenAI's First Unified Model, Truly Native * 63. Interview with Rokid's Misa Zhu: AI Glasses, Entry Points, and the Ecosystem Battle Amid Tech Giants * Practices and Reflections on Vibe Coding in Code Generation and Collaboration * Don't Let Format Kill Ideas: Logics-Parsing V2 Defines New Boundaries for Document Parsing HomeArticlesPodcastsVideosTweets