--- title: GLM-5.3-Flash | daily.dev description: GLM-5.3-Flash is a new natively multimodal model in the GLM-5 series from Z.ai, with 320B total parameters but only 18B active parameters via a... canonical: https://daily.dev/posts/glm-5-3-flash-bh9uwyinw twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: GLM-5.3-Flash | daily.dev og:description: GLM-5.3-Flash is a new natively multimodal model in the GLM-5 series from Z.ai, with 320B total parameters but only 18B active parameters via a... og:url: https://daily.dev/posts/glm-5-3-flash-bh9uwyinw og:image: https://api.daily.dev/og/posts/bh9UwYinw.png og:image:alt: GLM-5.3-Flash 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. # GLM-5.3-Flash **[nerdalytics](https://daily.dev/sources/jcwojcedqzpspqbvk73nc)** 路 [@nerdalytics](https://daily.dev/nerdalytics) 路 14 upvotes 路 4 comments ## Summary GLM-5.3-Flash is a new natively multimodal model in the GLM-5 series from Z.ai, with 320B total parameters but only 18B active parameters via a mixture-of-experts-style design. It introduces a hybrid sparse and linear attention architecture plus Manifold-Constrained Hyper-Connections (mHC), trained on a 30T-token multimodal corpus. It reportedly outperforms GLM-5.2 across benchmarks and real-world workloads at one-tenth the price, while approaching Claude Opus 4.8 on coding and agentic benchmarks. The model card includes deployment instructions for Transformers, vLLM, SGLang, KTransformers, and Docker, along with benchmark results on Terminal-Bench 2.1, DeepSWE, HLE, and other agentic evaluation suites. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Community discussion Top comments from developers on daily.dev. **@geniusatwork** 路 3 upvotes > Qwen3.8-Flash and GLM-5.3-Flash on the same day 馃 **@nerdalytics** 路 2 upvotes > Ox Alpha is quite cheap, but it's not a great model. > > > Started a worktree+branch, after compacting the session this model started to edit files in main branch. > > > I have a fork and workered on issue, I asked the model to check how far behind we are compared to origin/HEAD. It only checked my fork. Also ignored details like the repo URL of the upstream repo, and kept looking for CI/CD feedback on my fork. > > > It fumbled with Rust code in a Tauri+Svelte project, and kept calling my clippy settings pedantic and aggressive. The Svelte code wasn't great either. Pasting a screenshot with edited... **@agustinbarrientos** 路 0 upvotes > What peak VRAM and tokens per second did each official serving path reach? Did any path fit hardware that the active-parameter count made look unrealistic? ## Similar posts on daily.dev - [Don鈥檛 just attend KubeCon \+ CloudNativeCon, Merge Forward your experience\!](https://daily.dev/posts/don-t-just-attend-kubecon-cloudnativecon-merge-forward-your-experience--l0rpp73x8) 路 CNCF 路 0 upvotes 路 0 comments - [Announcing H2 2026 KCDs](https://daily.dev/posts/announcing-h2-2026-kcds-m96goajm1) 路 CNCF 路 1 upvotes 路 0 comments - [Two months of Open Community Groups](https://daily.dev/posts/two-months-of-open-community-groups-asf52zhbs) 路 CNCF 路 0 upvotes 路 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#deep-learning](https://daily.dev/tags/deep-learning), [#vllm](https://daily.dev/tags/vllm) [View this post on daily.dev](https://daily.dev/posts/glm-5-3-flash-bh9uwyinw) ```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":"DiscussionForumPosting","mainEntityOfPage":"https://daily.dev/posts/glm-5-3-flash-bh9uwyinw","headline":"GLM-5.3-Flash","text":"Shared: zai-org/GLM-5.3-Flash 路 Hugging Face","url":"https://daily.dev/posts/glm-5-3-flash-bh9uwyinw","datePublished":"2026-08-26T15:10:21.068Z","dateModified":"2026-08-26T15:48:03.393Z","author":{"@type":"Person","name":"nerdalytics","url":"https://daily.dev/nerdalytics","image":"https://avatars.githubusercontent.com/u/97166791?v=4","description":"Husband. Cat and dog dad. Nerd.","interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"EndorseAction"},"userInteractionCount":3510}},"image":"https://media.daily.dev/image/upload/s--OHB84bZF--/f_auto/v1722860399/public/Placeholder%2010","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":14},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":4}],"sharedContent":{"@type":"WebPage","url":"https://api.daily.dev/r/yPk0ntHqf"},"comment":[{"@type":"Comment","text":"Qwen3.8-Flash and GLM-5.3-Flash on the same day 馃","datePublished":"2026-08-26T15:24:32.054Z","url":"https://daily.dev/posts/bh9UwYinw#c-wMe7a7hJ6","author":{"@type":"Person","name":"Sam Dolin","url":"https://daily.dev/geniusatwork","image":"https://media.daily.dev/image/upload/s--nUhoZxKm--/f_auto/v1787434895/avatars/avatar_75QyPrL3BZgArSo6aFWuZ?_a=BAMAMicg0"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":3}},{"@type":"Comment","text":"Ox Alpha is quite cheap, but it鈥檚 not a great model.\nStarted a worktree+branch, after compacting the session this model started to edit files in main branch.\nI have a fork and workered on issue, I asked the model to check how far behind we are compared to origin/HEAD. It only checked my fork. Also ignored details like the repo URL of the upstream repo, and kept looking for CI/CD feedback on my fork.\nIt fumbled with Rust code in a Tauri+Svelte project, and kept calling my clippy settings pedantic and aggressive. The Svelte code wasn鈥檛 great either. Pasting a screenshot with edited highlights seems to confuse the model. Like I added a red box around the UI elements that need fixing, and it went looking where the in code the red borders|nofill-rectangles are. For some reason it believed I meant to get rid of the red boxes.\nBut it did also good. Checking out a big PR and reviewing it using my skills took an hour, but cost only $0.05 cents. It found some issues and managed to fix them.\nNot bad, but not impressive either. The price tag is amazing; 5 cents for a PR review.","datePublished":"2026-08-26T17:20:34.052Z","url":"https://daily.dev/posts/bh9UwYinw#c-XudRF3PDh","author":{"@type":"Person","name":"nerdalytics","url":"https://daily.dev/nerdalytics","image":"https://avatars.githubusercontent.com/u/97166791?v=4"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":3}},{"@type":"Comment","text":"What peak VRAM and tokens per second did each official serving path reach? Did any path fit hardware that the active-parameter count made look unrealistic?","datePublished":"2026-08-27T12:41:18.864Z","url":"https://daily.dev/posts/bh9UwYinw#c-PWLGb7ll2","author":{"@type":"Person","name":"Agustin Barrientos","url":"https://daily.dev/agustinbarrientos","image":"https://media.daily.dev/image/upload/s--5ayxQnqn--/f_auto/v1783272917/avatars/avatar_wQYYVe5Tbj0NJ7C7qPoa8?_a=BAMAMicg0"}}],"isPartOf":{"@type":"WebPage","url":"https://daily.dev/sources/jcwojcedqzpspqbvk73nc","name":"nerdalytics"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"nerdalytics","item":"https://daily.dev/sources/jcwojcedqzpspqbvk73nc"},{"@type":"ListItem","position":3,"name":"GLM-5.3-Flash"}]} ```