{"id":"77fe0e1a5366","page":"/c/77fe0e1a5366/","topic":"Google bench 2.1 vs 4.0 performance","publisher":"@SemiAnalysis_","official":false,"platform":"X","publishedAt":"2026-09-08T00:00:26.000Z","model":"Gemini 3","form":"table","metrics":{"views":469782,"likes":2648,"reposts":158,"quotes":80,"replies":111,"saves":541},"readAt":"2026-10-13T18:30:45.158Z","chartsInPost":1,"score":79,"scoreWeights":{"views":0.4,"likes":0.3,"eng":0.15,"save":0.15},"quadrant":"wide","ranks":{"score":{"industry":{"n":69,"of":1162},"company":{"n":1,"of":24}},"views":{"industry":{"n":157,"of":1162},"company":{"n":1,"of":24}},"likes":{"industry":{"n":110,"of":1162},"company":{"n":1,"of":24}},"eng":{"industry":{"n":675,"of":1162},"company":{"n":11,"of":24}},"save":{"industry":{"n":352,"of":1162},"company":{"n":11,"of":24}}},"percentiles":{"industry":{"views":0.8653184165232358,"likes":0.9053356282271945,"eng":0.4195352839931153,"save":0.6975043029259896},"company":{"views":0.9791666666666666,"likes":0.9791666666666666,"eng":0.5625,"save":0.5625}},"peers":{"industry":1162,"company":24,"companyName":"@semianalysis_"},"medians":{"industry":{"views":39945,"likes":289,"eng":0.007331788144862199,"save":0.0007642791742968587},"company":{"views":40346.5,"likes":212.5,"eng":0.0062466201307562995,"save":0.001023405708315584},"form":{"n":95,"views":133203,"eng":0.0060924409940925}},"ratios":{"vsCompanyViews":11.643686565129565,"vsCompanyLikes":12.461176470588235,"viewsPerFollower":null,"viewsPerDay":13133.020513037278},"followers":null,"reads":[{"at":"2026-10-12T15:08:42.451Z","views":469715,"likes":2649,"reposts":158,"quotes":80,"replies":111,"saves":541,"source":"x-search-cache"},{"at":"2026-10-12T15:18:03.931Z","views":469715,"likes":2649,"reposts":158,"quotes":80,"replies":111,"saves":541,"source":"x-search-cache"},{"at":"2026-10-12T20:13:56.150Z","views":469735,"likes":2649,"reposts":158,"quotes":80,"replies":111,"saves":541,"source":"x-search-cache"},{"at":"2026-10-12T20:17:56.286Z","views":469735,"likes":2650,"reposts":158,"quotes":80,"replies":111,"saves":541,"source":"x-search-cache"},{"at":"2026-10-12T20:54:59.888Z","views":469735,"likes":2650,"reposts":158,"quotes":80,"replies":111,"saves":541,"source":"x-search-cache"},{"at":"2026-10-13T18:30:45.158Z","views":469782,"likes":2648,"reposts":158,"quotes":80,"replies":111,"saves":541,"source":"bird-read"}],"chartData":"/api/charts/a99d6fdd4cda.json","copy":{"en":{"why":[{"factor":"reach-vs-company","text":"Its 470K views are 12x this company's median post, the #1 of 24 @SemiAnalysis_ posts by views."},{"factor":"rank","text":"A score of 79 puts it at #69 of 1,162 AI chart posts and #1 of 24 @SemiAnalysis_ posts."},{"factor":"launch-day","text":"It was the #1 of 2 posts from this account that day, against 31.5K views for the other."}],"verdict":"It sits in the Wide but thin quadrant, at #69 of 1,162 AI chart posts.","news":{"headline":"@SemiAnalysis_ table on Google bench 2.1 vs 4.0","dek":"A table chart on Google bench 2.1 vs 4.0 performance drew 470K views and a 0.64% engagement rate.","what":"On September 8, 2026, @SemiAnalysis_ posted a table chart on Google bench 2.1 vs 4.0 performance. That day the account posted 2 items, the other on DeepSWE v1.1 Model Cost vs Performance at 31.5K views.","numbers":"It reached 470K views, #157 of 1,162 AI chart posts, against a median of 39.9K views among AI chart posts. Its score of 79 ranks #69 of 1,162 AI chart posts.","lesson":"A table chart with 4 models shown drew 470K views but a 0.64% engagement rate, below the AI chart median. The chart shows Terminal-Bench 2.1 to 4.0 drops of -30.7% and -35.6%.","implication":"At #1 of 24 @SemiAnalysis_ posts by views, this table chart reached 12x the company median of 40.3K views. Its 0.64% engagement rate sits below the AI chart median, so the reach and the engagement do not match."}},"zh":{"why":[{"factor":"reach-vs-company","text":"浏览量47万，是公司中位数4万的12倍，排第1。"},{"factor":"rank","text":"得分79，在1162条AI图表帖中排第69，公司内排第1。"},{"factor":"launch-day","text":"发布当天发2条，本条浏览量排第1。"}],"verdict":"传得远，互动偏少，得分在1162条AI图表帖中排第69。","news":{"headline":"@SemiAnalysis_ 图表浏览量47万","dek":"TERMINAL-BENCH 2.1到4.0的得分下降","what":"2026年9月8日，@SemiAnalysis_ 发布1张图表，对比4个模型在TERMINAL-BENCH 2.1和4.0的得分。当天还有1条关于DeepSWE v1.1的帖子。","numbers":"浏览量47万，在1162条AI图表帖中排第157，公司24条帖子中排第1。互动率0.64%，收藏率0.12%。","lesson":"浏览量是公司中位数4万的12倍，但互动率低于中位数，属于传得远但互动偏少。","implication":"在@SemiAnalysis_的24条帖子中排第1，但互动率0.64%低于AI图表帖中位数，说明高浏览量未转化为高互动。"}},"lane":"en:amd/DeepSeek-V4.1-Flash zh:amd/DeepSeek-V4-Flash","checkedAt":"2026-10-13T20:59:20.888Z"}}