{"rewrite":{"id":"r_2296e02eb6802962345a229a","clusterId":"c_a172e5a77726f4d2888b7b9f","slug":"chinese-ai-model-glm-5-2-surpasses-claude-opus-4-7-in-benchmarks","model":"deepseek-v4-flash","headline":"Chinese AI Model GLM-5.2 Surpasses Claude Opus 4.7 in Benchmarks","summary":"Chinese AI company Z.ai officially announced GLM-5.2 on June 17, 2026. The 753-billion-parameter model surpasses Claude Opus 4.7 in multiple benchmarks, outperforms Claude Fable 5 in human blind tests for coding, and is available as an open model under the MIT License.","whyItMatters":"GLM-5.2 shows that Chinese open-source models continue to close the gap with top proprietary systems, matching or beating Anthropic's latest in specific coding and long-horizon tasks.","webCardHtml":"\u003cp\u003eZ.ai released GLM-5.2 on June 17, following a partial reveal on June 13 when Anthropic discontinued Claude Fable 5 and Claude Mythos 5 services. The model has 753 billion parameters and supports 1 million token input. In benchmark comparisons, GLM-5.2 surpassed GPT-5.5 in multiple tests and approached Claude Opus 4.8 scores. It excelled in long-duration task execution, beating both Claude Opus 4.7 and GPT-5.5. In the Arena coding performance test, GLM-5.2 ranked second worldwide, behind only Claude Opus 4.8. In the Design Arena coding test, it defeated Claude Fable 5 and took first place. The model is available on Z.ai\u0026#39;s chat service and can be used with coding agents including ZCode, Claude Code, and OpenCode. It is distributed as an open model on Hugging Face under the MIT License.\u003c/p\u003e","blueskyPost":"GLM-5.2's open MIT License means any developer can download and run a 753-billion-parameter model that beats Claude Opus 4.7. The closed vs. open AI debate just got a concrete data point from China.","twitterPost":"GLM-5.2 beats Claude Opus 4.7 and Fable 5 in coding tests. It's open source under MIT. The gap between open and closed models is narrowing fast.","threadsPost":"GLM-5.2 from Z.ai is a 753-billion-parameter open model that beats Claude Opus 4.7 across benchmarks and outperforms Claude Fable 5 in human-blind coding tests. Under MIT License, anyone can download it. The closed-model advantage is eroding, and this release from China makes that concrete.","newsletterBlurb":"Z.ai officially announced GLM-5.2, a 753-billion-parameter open model that surpasses Claude Opus 4.7 in benchmarks and beat Claude Fable 5 in a coding blind test. It is available under the MIT License on Hugging Face.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20260617-z-ai-glm-5-2-release/\",\"title\":\"Chinese Model \\\"GLM-5.2\\\" Surpassing Claude Opus 4.7 Officially Announced, Outperforms Claude Fable 5 in Some Tests and Available for Download as Open Model\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":4438,"outputTokens":638,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1781663158,"createdAt":"2026-06-17T02:16:17.000Z","publishedAt":"2026-06-17T02:19:57.000Z","updatedAt":"2026-06-17T02:19:57.000Z"},"cluster":{"id":"c_a172e5a77726f4d2888b7b9f","canonicalTitle":"Claude Opus 4.7超えの中華モデル「GLM-5.2」が正式発表される、一部テストではClaude Fable 5を上回りオープンモデルとして誰でもダウンロード可能に","representativeArticleId":"a_a3e2251908edb9408512d1b5","sourceCount":2,"writtenSourceCount":1,"writeAttempts":0,"isSolo":false,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[],\"studios\":[],\"people\":[],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-06-17T01:43:00.000Z","lastSeenAt":"2026-06-17T10:10:00.000Z","updatedAt":"2026-06-18T01:00:46.000Z"},"attribution":[{"source":"ASCII.jp","url":"https://ascii.jp/elem/000/004/411/4411485/?rss","title":"100万トークンの超長文コンテキスト対応モデル「Z.ai GLM-5.2」をOrcaRouterで提供開始 ― オープンモデルのコストでClaude Opus級のコーディング性能を実現"},{"source":"GIGAZINE","url":"https://gigazine.net/news/20260617-z-ai-glm-5-2-release/","title":"Claude Opus 4.7超えの中華モデル「GLM-5.2」が正式発表される、一部テストではClaude Fable 5を上回りオープンモデルとして誰でもダウンロード可能に"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":[],"studios":[],"people":[],"type":"news","domain":"other","is_roundup":false},"keyFacts":["Chinese AI company Z.ai released GLM-5.2 on June 17, 2026.","GLM-5.2 has 753 billion parameters and supports 1 million token input.","GLM-5.2 surpassed Claude Opus 4.7 in multiple benchmarks and outperformed Claude Fable 5 in human blind tests for coding.","In the Design Arena coding test, GLM-5.2 defeated Claude Fable 5 and took first place.","GLM-5.2 is available as an open model under the MIT License on Hugging Face."]}