--- title: Alibaba takes aim at OpenAI and Anthropic with... description: Alibaba launched Qwen3.8-Max, a 2.4-trillion-parameter mixture-of-experts model that activates ~95 billion parameters per inference. The open-weight model... canonical: https://daily.dev/posts/alibaba-takes-aim-at-openai-and-anthropic-with-qwen3-8-max-launch-xvlfrlsit twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Alibaba takes aim at OpenAI and Anthropic with Qwen3.8-Max launch | daily.dev og:description: Alibaba launched Qwen3.8-Max, a 2.4-trillion-parameter mixture-of-experts model that activates ~95 billion parameters per inference. The open-weight model... og:url: https://daily.dev/posts/alibaba-takes-aim-at-openai-and-anthropic-with-qwen3-8-max-launch-xvlfrlsit og:image: https://api.daily.dev/og/posts/XvLFRLsIt.png og:image:alt: Alibaba takes aim at OpenAI and Anthropic with Qwen3.8-Max launch og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Alibaba takes aim at OpenAI and Anthropic with Qwen3.8-Max launch **[InfoWorld](https://daily.dev/sources/infoworld)** · 5 min read · 1 upvotes · 0 comments ## Summary Alibaba launched Qwen3.8-Max, a 2.4-trillion-parameter mixture-of-experts model that activates ~95 billion parameters per inference. The open-weight model targets enterprise software engineering, multimodal reasoning, and knowledge-intensive workloads, with benchmarks comparing it against Claude Opus 4.8 and GPT-5.6 Sol on coding tasks like SWE-bench Pro. Alibaba claims the model autonomously completed a coding project over 16 days, though analysts urge scrutiny of that claim. Industry analysts highlight that inference efficiency and open-weight availability are increasingly important to enterprises, but caution about data sovereignty concerns, lack of indemnification protections, and the gap between publishing weights and true open-source availability. The smaller Qwen3.8-27B variant may be more practically deployable for most organizations. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is Qwen3.8-Max and how many parameters does it activate during inference? Qwen3.8-Max is Alibaba's largest AI model, a 2.4-trillion-parameter mixture-of-experts model that activates only about 95 billion parameters per request during inference. This design is intended to lower inference costs while supporting coding, reasoning, and multimodal tasks. Open-weight versions were scheduled for release through Alibaba Cloud's Model Studio the week following the announcement. _Engineers weighing model efficiency against cost can track releases like this one on daily.dev._ ### How does Qwen3.8-Max compare to Claude Opus 4.8 and GPT-5.6 Sol on coding benchmarks? Alibaba published internal test results comparing Qwen3.8-Max against Claude Opus 4.8, Claude Fable 5, and OpenAI's GPT-5.6 Sol on coding benchmarks including SWE-bench Pro and a proprietary NL2Repo-Bench evaluation, using each vendor's own coding harness (Claude Code and Codex) and reporting the highest published score for each rival. Analysts noted these self-reported comparisons deserve scrutiny before drawing conclusions. _Developers comparing coding models before committing can follow benchmark disputes like this on daily.dev._ ### Why might enterprises hesitate to deploy Qwen3.8-Max in production despite its open-weight release? Enterprises outside China may be wary of relying on models hosted within China, often opting instead for hyperscaler or on-premises deployment, both of which add cost. Open-weight models also typically lack the indemnification protections that commercial vendors provide, so organizations need their own security, governance, and code-scanning controls to catch copyright and intellectual property risks before production use. _Teams weighing sovereignty and licensing risk in open-weight models can follow these tradeoffs on daily.dev._ ## Similar posts on daily.dev - [Alibaba’s Qwen3-Max-Thinking expands enterprise AI model choices](https://daily.dev/posts/alibaba-s-qwen3-max-thinking-expands-enterprise-ai-model-choices-nv9ulemja) · InfoWorld · 0 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-coding](https://daily.dev/tags/ai-coding), [#qwen](https://daily.dev/tags/qwen), [#alibaba](https://daily.dev/tags/alibaba), [#mixture-of-experts](https://daily.dev/tags/mixture-of-experts) [View this post on daily.dev](https://daily.dev/posts/alibaba-takes-aim-at-openai-and-anthropic-with-qwen3-8-max-launch-xvlfrlsit) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Alibaba takes aim at OpenAI and Anthropic with Qwen3.8-Max launch","url":"https://daily.dev/posts/alibaba-takes-aim-at-openai-and-anthropic-with-qwen3-8-max-launch-xvlfrlsit","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/alibaba-takes-aim-at-openai-and-anthropic-with-qwen3-8-max-launch-xvlfrlsit"},"datePublished":"2026-08-03T12:29:54.486Z","dateModified":"2026-09-14T08:31:17.949Z","description":"Alibaba launched Qwen3.8-Max, a 2.4-trillion-parameter mixture-of-experts model that activates ~95 billion parameters per inference. The open-weight model...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/e7863c91cf42e77cf93b5d2f71603050?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/e7863c91cf42e77cf93b5d2f71603050?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"InfoWorld","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"InfoWorld","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/bf6d68a999064029b0bb09aa6268f1f3","url":"https://daily.dev/sources/infoworld"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/alibaba-takes-aim-at-openai-and-anthropic-with-qwen3-8-max-launch-xvlfrlsit","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,ai-coding,qwen,alibaba,mixture-of-experts","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"InfoWorld","item":"https://daily.dev/sources/infoworld"},{"@type":"ListItem","position":3,"name":"Alibaba takes aim at OpenAI and Anthropic with Qwen3.8-Max launch"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/alibaba-takes-aim-at-openai-and-anthropic-with-qwen3-8-max-launch-xvlfrlsit#faq","mainEntity":[{"@type":"Question","name":"What is Qwen3.8-Max and how many parameters does it activate during inference?","acceptedAnswer":{"@type":"Answer","text":"Qwen3.8-Max is Alibaba's largest AI model, a 2.4-trillion-parameter mixture-of-experts model that activates only about 95 billion parameters per request during inference. This design is intended to lower inference costs while supporting coding, reasoning, and multimodal tasks. Open-weight versions were scheduled for release through Alibaba Cloud's Model Studio the week following the announcement. Engineers weighing model efficiency against cost can track releases like this one on daily.dev."}},{"@type":"Question","name":"How does Qwen3.8-Max compare to Claude Opus 4.8 and GPT-5.6 Sol on coding benchmarks?","acceptedAnswer":{"@type":"Answer","text":"Alibaba published internal test results comparing Qwen3.8-Max against Claude Opus 4.8, Claude Fable 5, and OpenAI's GPT-5.6 Sol on coding benchmarks including SWE-bench Pro and a proprietary NL2Repo-Bench evaluation, using each vendor's own coding harness (Claude Code and Codex) and reporting the highest published score for each rival. Analysts noted these self-reported comparisons deserve scrutiny before drawing conclusions. Developers comparing coding models before committing can follow benchmark disputes like this on daily.dev."}},{"@type":"Question","name":"Why might enterprises hesitate to deploy Qwen3.8-Max in production despite its open-weight release?","acceptedAnswer":{"@type":"Answer","text":"Enterprises outside China may be wary of relying on models hosted within China, often opting instead for hyperscaler or on-premises deployment, both of which add cost. Open-weight models also typically lack the indemnification protections that commercial vendors provide, so organizations need their own security, governance, and code-scanning controls to catch copyright and intellectual property risks before production use. Teams weighing sovereignty and licensing risk in open-weight models can follow these tradeoffs on daily.dev."}}]} ```