--- title: I run a local coding model on my 8GB GPU for real UI... description: A developer describes running a local coding model (Qwen 3.5 9B, Q4_K_M quantization) on an 8GB VRAM GPU as a replacement for Claude Design's UI generation... canonical: https://daily.dev/posts/i-run-a-local-coding-model-on-my-8gb-gpu-for-real-ui-work-and-i-m-not-going-back-to-claude-design-iqcjufmir twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: I run a local coding model on my 8GB GPU for real UI work, and I'm not going back to Claude Design | daily.dev og:description: A developer describes running a local coding model (Qwen 3.5 9B, Q4_K_M quantization) on an 8GB VRAM GPU as a replacement for Claude Design's UI generation... og:url: https://daily.dev/posts/i-run-a-local-coding-model-on-my-8gb-gpu-for-real-ui-work-and-i-m-not-going-back-to-claude-design-iqcjufmir og:image: https://api.daily.dev/og/posts/iqcjufMir.png og:image:alt: I run a local coding model on my 8GB GPU for real UI work, and I'm not going back to Claude Design 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. # I run a local coding model on my 8GB GPU for real UI work, and I'm not going back to Claude Design **[XDA Developers](https://daily.dev/sources/xda-developers)** · 5 min read · 0 upvotes · 0 comments ## Summary A developer describes running a local coding model (Qwen 3.5 9B, Q4_K_M quantization) on an 8GB VRAM GPU as a replacement for Claude Design's UI generation workflow. The setup pairs the local model with tools like Jan, LM Studio, VS Code Live Server, and Open Design, noting trade-offs like weaker first-try polish but the ability to run many more iterations without a credit cap, and full offline operation. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Can an 8GB VRAM GPU run a local coding model for UI vibe coding work? Yes, an 8GB VRAM GPU can run Qwen 3.5 9B at Q4_K_M quantization, which takes about 5GB, leaving headroom depending on context length and KV cache size. Larger models like Qwen 3.6 35B-A3B at Q2_K_XL were too large to run on 8GB VRAM even at aggressive quantization. _daily.dev surfaces hands-on comparisons for developers sizing local models to their own hardware limits._ ### What are the trade-offs of using a local LLM instead of Claude Design for UI design work? Local models produce less polished first-try results than Claude's Opus models and lack the 'second opinion' creative input bigger cloud models sometimes offer. However, local setups allow running roughly 40 iterations in the time and token budget that Claude uses for five, with no credit cap and full offline operation, making iteration-heavy workflows cheaper overall. _developers weighing local versus cloud AI design tools can track these comparisons on daily.dev._ ## Similar posts on daily.dev - [My 8GB GPU shouldn't run flagship local LLMs, but this workflow makes it work anyway](https://daily.dev/posts/my-8gb-gpu-shouldn-t-run-flagship-local-llms-but-this-workflow-makes-it-work-anyway-nb2r0gq97) · XDA Developers · 1 upvotes · 0 comments - [I replaced Claude Pro with a local 9B model for a week, and finally found out what I was paying $20 a month for](https://daily.dev/posts/i-replaced-claude-pro-with-a-local-9b-model-for-a-week-and-finally-found-out-what-i-was-paying-20--hoed7m5zq) · XDA Developers · 0 upvotes · 0 comments - [The Best Local Coding Models for Any Setup](https://daily.dev/posts/the-best-local-coding-models-for-any-setup-8cm3jivdh) · Kilo Blog · 1 upvotes · 0 comments --- Tags: [#gpu](https://daily.dev/tags/gpu), [#claude](https://daily.dev/tags/claude), [#vibe-coding](https://daily.dev/tags/vibe-coding), [#local-ai](https://daily.dev/tags/local-ai), [#qwen](https://daily.dev/tags/qwen) [View this post on daily.dev](https://daily.dev/posts/i-run-a-local-coding-model-on-my-8gb-gpu-for-real-ui-work-and-i-m-not-going-back-to-claude-design-iqcjufmir) ```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":"I run a local coding model on my 8GB GPU for real UI work, and I'm not going back to Claude Design","url":"https://daily.dev/posts/i-run-a-local-coding-model-on-my-8gb-gpu-for-real-ui-work-and-i-m-not-going-back-to-claude-design-iqcjufmir","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/i-run-a-local-coding-model-on-my-8gb-gpu-for-real-ui-work-and-i-m-not-going-back-to-claude-design-iqcjufmir"},"datePublished":"2026-08-17T18:04:12.046Z","dateModified":"2026-08-17T18:32:45.746Z","description":"A developer describes running a local coding model (Qwen 3.5 9B, Q4_K_M quantization) on an 8GB VRAM GPU as a replacement for Claude Design's UI generation...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/d225da968285b8d33789fb16f1126681?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/d225da968285b8d33789fb16f1126681?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"XDA Developers","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":"XDA Developers","logo":"https://media.daily.dev/image/upload/s--kCg6yyAP--/f_auto,q_auto/v1774964407/logos/xda-developers?_a=BAMAMiWQ0","url":"https://daily.dev/sources/xda-developers"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/i-run-a-local-coding-model-on-my-8gb-gpu-for-real-ui-work-and-i-m-not-going-back-to-claude-design-iqcjufmir","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"gpu,claude,vibe-coding,local-ai,qwen","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"XDA Developers","item":"https://daily.dev/sources/xda-developers"},{"@type":"ListItem","position":3,"name":"I run a local coding model on my 8GB GPU for real UI work, and I'm not going back to Claude Design"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/i-run-a-local-coding-model-on-my-8gb-gpu-for-real-ui-work-and-i-m-not-going-back-to-claude-design-iqcjufmir#faq","mainEntity":[{"@type":"Question","name":"Can an 8GB VRAM GPU run a local coding model for UI vibe coding work?","acceptedAnswer":{"@type":"Answer","text":"Yes, an 8GB VRAM GPU can run Qwen 3.5 9B at Q4_K_M quantization, which takes about 5GB, leaving headroom depending on context length and KV cache size. Larger models like Qwen 3.6 35B-A3B at Q2_K_XL were too large to run on 8GB VRAM even at aggressive quantization. daily.dev surfaces hands-on comparisons for developers sizing local models to their own hardware limits."}},{"@type":"Question","name":"What are the trade-offs of using a local LLM instead of Claude Design for UI design work?","acceptedAnswer":{"@type":"Answer","text":"Local models produce less polished first-try results than Claude's Opus models and lack the 'second opinion' creative input bigger cloud models sometimes offer. However, local setups allow running roughly 40 iterations in the time and token budget that Claude uses for five, with no credit cap and full offline operation, making iteration-heavy workflows cheaper overall. developers weighing local versus cloud AI design tools can track these comparisons on daily.dev."}}]} ```