{"rewrite":{"id":"r_b6776952dc90e62ae05062d7","clusterId":"c_03795a884b07693e45b8c354","slug":"grok-4-6-arrives-with-long-duration-agent-focus-matching-gpt-5-6-sol","model":"deepseek-v4-flash:free","headline":"Grok 4.6 Arrives With Long-Duration Agent Focus, Matching GPT-5.6 Sol","summary":"SpaceXAI announced Grok 4.6 on August 12, 2026, an upgrade to Grok 4.5 at the same price. It scores 61 on the Artificial Intelligence Analysis Index, matching GPT-5.6 Sol, and shows notable gains in long-running agent tasks and coding benchmarks. The model emphasizes task continuation and converting vague ideas into working applications.","whyItMatters":"Grok 4.6 brings SpaceXAI back to the intelligence frontier alongside OpenAI, positioned just behind Anthropic, per Artificial Analysis.","webCardHtml":"\u003cp\u003eSpaceXAI\u0026#39;s Grok 4.6, announced August 12, 2026, is built on Grok 4.5 with a focus on long-duration agent tasks and interactive visual processing. On the Artificial Intelligence Analysis Index, it scored 61 points, matching OpenAI\u0026#39;s GPT-5.6 Sol. The model recorded strong results on coding benchmarks: 1753 on GDPVal-AA v2, 69.9% on CursorBench 3.2, and 61.3% on FrontierCode v1.1, second only to Claude Fable 5 Max.\u003c/p\u003e\u003cp\u003eTraining involved additional learning over a longer period than Grok 4.5, with regenerated SFT trajectories and model-based filtering. SpaceXAI highlights task continuation capability, positioning the model to turn vague product ideas into working first versions. Safety measures were strengthened, with extensive pre-deployment testing.\u003c/p\u003e","blueskyPost":"Grok 4.6 is out, matching GPT-5.6 Sol on the AAII at 61 points. SpaceXAI says it's strongest at long agent tasks and coding, with a focus on turning vague ideas into working apps.","twitterPost":"Grok 4.6 launches, scoring 61 on the AAII, matching GPT-5.6 Sol. SpaceXAI highlights long-duration agent tasks and coding, with gains over Grok 4.5.","threadsPost":null,"newsletterBlurb":"SpaceXAI released Grok 4.6, an upgrade that matches GPT-5.6 Sol on the AAII at 61 points. The model focuses on long-running agent tasks and coding, with strong results on several benchmarks. Artificial Analysis says SpaceXAI is back at the frontier, just behind Anthropic.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20260813-grok-4-6/\",\"title\":\"Grok 4.6 arrives, stronger at long-duration agent tasks, matching GPT-5.6 Sol and Claude Fable 5 in performance\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":4646,"outputTokens":617,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1786780199,"createdAt":"2026-08-15T07:44:58.000Z","publishedAt":"2026-08-15T07:46:44.000Z","updatedAt":"2026-08-15T07:44:58.000Z"},"cluster":{"id":"c_03795a884b07693e45b8c354","canonicalTitle":"「Grok 4.6」が登場、長時間のエージェント作業に強くなりGPT-5.6 SolやClaude Fable 5と並ぶ性能に","representativeArticleId":"a_f889a3ecf08d21569e6508ab","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"Grok 4.6\"],\"studios\":[],\"people\":[],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-08-13T04:33:00.000Z","lastSeenAt":"2026-08-13T04:33:00.000Z","updatedAt":"2026-08-15T07:46:44.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20260813-grok-4-6/","title":"「Grok 4.6」が登場、長時間のエージェント作業に強くなりGPT-5.6 SolやClaude Fable 5と並ぶ性能に"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["Grok 4.6"],"studios":[],"people":[],"type":"news","domain":"other","is_roundup":false},"keyFacts":null}