--- title: Chinese AI company Zhipu claims its new is a better... description: Chinese AI firm Zhipu released GLM-5.3, claiming it outperforms Anthropic's and OpenAI's models on CyberGym, a benchmark for real-world cybersecurity... canonical: https://daily.dev/posts/chinese-ai-company-zhipu-claims-its-new-is-a-better-bug-finder-than-anthropic-openai-dwslwaplh twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Chinese AI company Zhipu claims its new is a better bug-finder than Anthropic, OpenAI | daily.dev og:description: Chinese AI firm Zhipu released GLM-5.3, claiming it outperforms Anthropic's and OpenAI's models on CyberGym, a benchmark for real-world cybersecurity... og:url: https://daily.dev/posts/chinese-ai-company-zhipu-claims-its-new-is-a-better-bug-finder-than-anthropic-openai-dwslwaplh og:image: https://api.daily.dev/og/posts/DWsLwApLH.png og:image:alt: Chinese AI company Zhipu claims its new is a better bug-finder than Anthropic, OpenAI 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. # Chinese AI company Zhipu claims its new is a better bug-finder than Anthropic, OpenAI **[The Register](https://daily.dev/sources/theregister)** · 6 min read · 0 upvotes · 0 comments ## Summary Chinese AI firm Zhipu released GLM-5.3, claiming it outperforms Anthropic's and OpenAI's models on CyberGym, a benchmark for real-world cybersecurity vulnerability discovery and exploitation chain reasoning. Testing on real codebases reportedly found 2,436 vulnerabilities across 269 projects, including 1,097 medium-to-high severity issues, some decades old. The model lagged behind Western rivals on other coding and security benchmarks, but the result suggests China is closing the gap in offensive security capability shortly after Anthropic's Mythos debut. Also covered: South Korea moving toward sanctions against Apple and Google over app store payment fees, India banning pre-ride tipping prompts for rideshare apps, TCS and HCL disclosing employee data breaches, Lenovo's infrastructure unit posting record profits, and India's ISRO tendering an astronaut training facility. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How does Zhipu's GLM-5.3 model compare to Anthropic and OpenAI models at finding security vulnerabilities? GLM-5.3 reportedly beats Fable 5 and GPT-5.6 Sol on the CyberGym benchmark, which tests real-world cybersecurity vulnerability discovery and exploitation chain reasoning. Zhipu says the model can reason across multiple stages of an exploit rather than just spotting isolated flaws. However, GLM-5.3 performed worse than Western models on other security and coding benchmarks, so its edge is narrow and benchmark-specific. _Developers evaluating AI models for security tooling can track benchmark shifts like this on daily.dev._ ### How many vulnerabilities did Zhipu's GLM-5.3 model find when tested on real-world codebases? GLM-5.3 was used to scan real codebases at Chinese companies and found 2,436 vulnerabilities across 269 projects, including 1,097 rated medium-to-high severity. The flaws spanned system kernels, operating systems, browser engines, open-source infrastructure, web applications, and network protocols, with some bugs dating back roughly 40 years. _Teams assessing AI-driven vulnerability scanning tools can follow real-world results like these on daily.dev._ ## Similar posts on daily.dev - [Zhipu share sale: China's AI star raises billions](https://daily.dev/posts/zhipu-share-sale-china-s-ai-star-raises-billions-lo6oftdth) · The Next Web · 0 upvotes · 0 comments - [Chinese cybersecurity company claims it’s built a better-than-Mythos bug finder](https://daily.dev/posts/chinese-cybersecurity-company-claims-it-s-built-a-better-than-mythos-bug-finder-00simhyab) · The Register · 0 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#vulnerability](https://daily.dev/tags/vulnerability), [#data-breach](https://daily.dev/tags/data-breach), [#ai-security](https://daily.dev/tags/ai-security) [View this post on daily.dev](https://daily.dev/posts/chinese-ai-company-zhipu-claims-its-new-is-a-better-bug-finder-than-anthropic-openai-dwslwaplh) ```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":"Chinese AI company Zhipu claims its new is a better bug-finder than Anthropic, OpenAI","url":"https://daily.dev/posts/chinese-ai-company-zhipu-claims-its-new-is-a-better-bug-finder-than-anthropic-openai-dwslwaplh","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/chinese-ai-company-zhipu-claims-its-new-is-a-better-bug-finder-than-anthropic-openai-dwslwaplh"},"datePublished":"2026-08-17T00:54:04.352Z","dateModified":"2026-08-18T11:05:05.007Z","description":"Chinese AI firm Zhipu released GLM-5.3, claiming it outperforms Anthropic's and OpenAI's models on CyberGym, a benchmark for real-world cybersecurity...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/8f0de59c794e2739e9e4451051d240cf?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/8f0de59c794e2739e9e4451051d240cf?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The Register","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":"The Register","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/66aa2113fdad463992ffcbf0e8963fda","url":"https://daily.dev/sources/theregister"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/chinese-ai-company-zhipu-claims-its-new-is-a-better-bug-finder-than-anthropic-openai-dwslwaplh","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,vulnerability,data-breach,ai-security","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The Register","item":"https://daily.dev/sources/theregister"},{"@type":"ListItem","position":3,"name":"Chinese AI company Zhipu claims its new is a better bug-finder than Anthropic, OpenAI"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/chinese-ai-company-zhipu-claims-its-new-is-a-better-bug-finder-than-anthropic-openai-dwslwaplh#faq","mainEntity":[{"@type":"Question","name":"How does Zhipu's GLM-5.3 model compare to Anthropic and OpenAI models at finding security vulnerabilities?","acceptedAnswer":{"@type":"Answer","text":"GLM-5.3 reportedly beats Fable 5 and GPT-5.6 Sol on the CyberGym benchmark, which tests real-world cybersecurity vulnerability discovery and exploitation chain reasoning. Zhipu says the model can reason across multiple stages of an exploit rather than just spotting isolated flaws. However, GLM-5.3 performed worse than Western models on other security and coding benchmarks, so its edge is narrow and benchmark-specific. Developers evaluating AI models for security tooling can track benchmark shifts like this on daily.dev."}},{"@type":"Question","name":"How many vulnerabilities did Zhipu's GLM-5.3 model find when tested on real-world codebases?","acceptedAnswer":{"@type":"Answer","text":"GLM-5.3 was used to scan real codebases at Chinese companies and found 2,436 vulnerabilities across 269 projects, including 1,097 rated medium-to-high severity. The flaws spanned system kernels, operating systems, browser engines, open-source infrastructure, web applications, and network protocols, with some bugs dating back roughly 40 years. Teams assessing AI-driven vulnerability scanning tools can follow real-world results like these on daily.dev."}}]} ```