{"rewrite":{"id":"r_805cbe8bab0bf0357e43088a","clusterId":"c_84ab2e6a136189366433eb8f","slug":"qwen3-8-27b-ties-gpt-5-6-luna-in-intelligence-beats-gpt-5-6-terra-in-agent-index","model":"deepseek-v4-flash:free","headline":"Qwen3.8 27B Ties GPT-5.6 Luna in Intelligence, Beats GPT-5.6 Terra in Agent Index","summary":"Alibaba released the open model Qwen3.8 27B on August 15. Third-party analysis from Artificial Analysis on August 17 gives it an Intelligence Index of 52, tying GPT-5.6 Luna, and an Agentic Index of 51, surpassing GPT-5.6 Terra and Claude Opus 4.8. The model runs locally.","whyItMatters":"A locally runnable open model now matches or beats several closed frontier models on third-party benchmarks, with no API fees.","webCardHtml":"\u003cp\u003eArtificial Analysis published its performance breakdown for Qwen3.8 27B on August 17, two days after Alibaba released the open model. The Intelligence Index score of 52 ties GPT-5.6 Luna, while the Agentic Index of 51 clears GPT-5.6 Terra and Claude Opus 4.8.\u003c/p\u003e\u003cp\u003eThe model is notable for its output volume. It generated 160 million tokens during the Intelligence Index test, against a median of 43 million. That verbosity raises context window and memory pressure, but for local users it avoids the API costs that high output would normally incur.\u003c/p\u003e","blueskyPost":"Qwen3.8 27B, a locally runnable open model, scores 52 on Artificial Analysis' Intelligence Index, tying GPT-5.6 Luna, and 51 on the Agentic Index, beating GPT-5.6 Terra.","twitterPost":"Qwen3.8 27B, a free open model that runs locally, ties GPT-5.6 Luna on intelligence (52) and beats GPT-5.6 Terra on agent performance (51), per Artificial Analysis.","threadsPost":null,"newsletterBlurb":"Alibaba's Qwen3.8 27B, a free open model that runs on home PCs, matches GPT-5.6 Luna on intelligence and beats GPT-5.6 Terra on agent performance in Artificial Analysis tests. Its high output token count is a cost advantage locally but strains context and memory.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20260818-qwen3-8-27b-performance/\",\"title\":\"Analysis: Locally Running Chinese Open Model 'Qwen3.8 27B' Exceeds GPT-5.6 Terra in Agent Performance, Significantly Outperforming Models of Similar Size\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":4751,"outputTokens":576,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1787050564,"createdAt":"2026-08-18T10:48:00.000Z","publishedAt":"2026-08-18T10:51:44.000Z","updatedAt":"2026-08-18T10:48:00.000Z"},"cluster":{"id":"c_84ab2e6a136189366433eb8f","canonicalTitle":"ローカルで動く中国製オープンモデル「Qwen3.8 27B」はGPT-5.6 Terra超えのエージェント性能で同等規模のモデルより圧倒的に高性能という分析結果","representativeArticleId":"a_df78be4dc718402d7cb5af93","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"Qwen3.8 27B\"],\"studios\":[\"Alibaba\"],\"people\":[],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-08-18T02:48:00.000Z","lastSeenAt":"2026-08-18T02:48:00.000Z","updatedAt":"2026-08-18T10:51:45.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20260818-qwen3-8-27b-performance/","title":"ローカルで動く中国製オープンモデル「Qwen3.8 27B」はGPT-5.6 Terra超えのエージェント性能で同等規模のモデルより圧倒的に高性能という分析結果"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["Qwen3.8 27B"],"studios":["Alibaba"],"people":[],"type":"news","domain":"other","is_roundup":false},"keyFacts":null}