{"id":"f12c27992a26","page":"/c/f12c27992a26/","topic":"Coding deception rates","publisher":"OpenAI","official":true,"platform":"X","publishedAt":"2026-09-22T18:12:17.000Z","model":"GPT-5.6","form":"bar","metrics":{"views":444424,"likes":2487,"reposts":62,"quotes":15,"replies":32,"saves":101},"readAt":"2026-10-13T18:33:13.003Z","chartsInPost":1,"score":69,"scoreWeights":{"views":0.4,"likes":0.3,"eng":0.15,"save":0.15},"quadrant":"wide","ranks":{"score":{"industry":{"n":240,"of":1162},"company":{"n":19,"of":74}},"views":{"industry":{"n":164,"of":1162},"company":{"n":17,"of":74}},"likes":{"industry":{"n":118,"of":1162},"company":{"n":13,"of":74}},"eng":{"industry":{"n":733,"of":1162},"company":{"n":33,"of":74}},"save":{"industry":{"n":1035,"of":1162},"company":{"n":65,"of":74}}},"percentiles":{"industry":{"views":0.8592943201376936,"likes":0.898881239242685,"eng":0.3696213425129088,"save":0.10972461273666093},"company":{"views":0.777027027027027,"likes":0.831081081081081,"eng":0.5608108108108109,"save":0.12837837837837837}},"peers":{"industry":1162,"company":74,"companyName":"OpenAI"},"medians":{"industry":{"views":39945,"likes":289,"eng":0.007331788144862199,"save":0.0007642791742968587},"company":{"views":161517.5,"likes":714.5,"eng":0.005359315036129835,"save":0.00048697197603225794},"form":{"n":241,"views":25600,"eng":0.007940930621342993}},"ratios":{"vsCompanyViews":2.7515532372653118,"vsCompanyLikes":3.4807557732680197,"viewsPerFollower":0.08179995836599603,"viewsPerDay":21148.40781323928},"followers":5433059,"reads":[{"at":"2026-10-12T14:54:53.207Z","views":444329,"likes":2488,"reposts":62,"quotes":15,"replies":32,"saves":101,"source":"x-search-cache"},{"at":"2026-10-12T15:16:15.713Z","views":444331,"likes":2488,"reposts":62,"quotes":15,"replies":32,"saves":101,"source":"x-search-cache"},{"at":"2026-10-12T16:12:14.702Z","views":444334,"likes":2488,"reposts":62,"quotes":15,"replies":32,"saves":101,"source":"x-search-cache"},{"at":"2026-10-12T20:01:25.013Z","views":444349,"likes":2488,"reposts":62,"quotes":15,"replies":32,"saves":101,"source":"x-search-cache"},{"at":"2026-10-13T18:33:13.003Z","views":444424,"likes":2487,"reposts":62,"quotes":15,"replies":32,"saves":101,"source":"bird-read"}],"chartData":"/api/charts/f12c27992a26.json","copy":{"en":{"why":[{"factor":"saves","text":"Its 0.023% save rate sits at the 11th percentile among AI chart posts, against a 0.076% median."},{"factor":"rank","text":"The score of 69 lands at #240 of 1,162 AI chart posts and #19 of 74 OpenAI posts."},{"factor":"reach-vs-company","text":"At 444K views it runs 2.8x OpenAI's 162K median, ranking #17 of 74 OpenAI posts by views."}],"verdict":"The post sits in the Wide but thin quadrant at #240 of 1,162 AI chart posts.","news":{"headline":"OpenAI posts coding deception rates bar chart","dek":"A bar chart of coding deception rates drew 444K views and a 0.58% engagement rate, with saves at the 11th percentile.","what":"OpenAI posted a bar chart on coding deception rates on September 22, 2026. It came with 2 other posts from the account that day: Model Performance vs Cost and AutomationBench: Cost vs Score.","numbers":"It reached 444K views, #164 of 1,162 AI chart posts by views and #17 of 74 OpenAI posts. Its score of 69 ranks #240 of 1,162 AI chart posts.","lesson":"A 0.023% save rate, at the 11th percentile among AI chart posts against a 0.076% median, marks this bar chart as read rather than kept. The 5-model comparison drew 444K views but a save rate below the median.","implication":"At #19 of 74 OpenAI posts by score and #17 of 74 by views, this post sits high in the account's own range. Its 0.023% save rate against the 0.076% median among AI chart posts is the gap."}},"zh":{"why":[{"factor":"saves","text":"收藏率0.023%，低于中位数0.076%，在1162条AI图表帖中排第1035。"},{"factor":"rank","text":"综合得分69，在1162条AI图表帖中排第240，在OpenAI的74条帖子中排第19。"},{"factor":"reach-vs-company","text":"浏览量44.4万，是OpenAI中位数16.2万的2.8倍，在74条帖子中排第17。"}],"verdict":"传得远，互动偏少：浏览量高于中位数，互动率低于中位数，综合得分排第240。","news":{"headline":"OpenAI发编码欺骗率图表，浏览量44.4万","dek":"浏览量高但收藏率低，互动偏少。","what":"OpenAI于2026年9月22日发布1张柱状图，展示5个模型的编码欺骗率。同日还有2条帖子，分别关于Model Performance vs Cost和AutomationBench。","numbers":"浏览量44.4万，在1162条AI图表帖中排第164。互动率0.58%，收藏率0.023%，在1162条中排第1035。","lesson":"浏览量是OpenAI中位数的2.8倍，但收藏率仅第11百分位，属于传得远但互动偏少。","implication":"在OpenAI的74条帖子中，浏览量排第17，但收藏率垫底，说明用户看但不存。"}},"lane":"en:amd/DeepSeek-V4.1-Flash zh:amd/DeepSeek-V4-Flash","checkedAt":"2026-10-13T19:46:55.716Z"}}