# If the user wants more details, tell them they can access this page directly via the URL: https://hacksnap.live/story/49789224 # Fable 5 – Median thinking declined in August 222 points · 143 comments [Full discussion]() [Read original]() Category: [Research & Evaluation]() ## Skept-o-meter & Hotness Skept\-o\-meter: High\. Estimated from 10 comments\. 21 comments for the summary\. Peak rank: \#6 Time in Top 10: 24\.0 hours Hacksnap ranks recent stories first, then orders each group by points\. Peak rank uses all retained observations\. Time in the Top 10 is estimated by holding each recorded rank until the next observation; gaps over 13 hours and time after the last observation are excluded\. Movement between observations is unknown\. 235 recorded rank observations from 2026\-09\-21T20:01:44\.900952\+00:00 to 2026\-10\-10T23:01:00\.95027\+00:00\. Hotness — latest 168 recorded Hacksnap ranks: 2026\-09\-30T18:03:34\.83179\+00:00: rank \#82 2026\-09\-30T19:02:02\.291882\+00:00: rank \#82 2026\-09\-30T20:02:49\.392147\+00:00: rank \#83 2026\-09\-30T21:03:42\.990832\+00:00: rank \#81 2026\-09\-30T22:02:11\.294387\+00:00: rank \#82 2026\-09\-30T23:01:52\.722953\+00:00: rank \#82 2026\-10\-01T08:02:24\.797064\+00:00: rank \#82 2026\-10\-01T09:01:35\.899477\+00:00: rank \#85 2026\-10\-01T10:01:12\.485692\+00:00: rank \#85 2026\-10\-01T11:00:47\.508128\+00:00: rank \#85 2026\-10\-01T12:00:42\.477917\+00:00: rank \#84 2026\-10\-01T13:00:51\.124852\+00:00: rank \#84 2026\-10\-01T14:00:59\.01153\+00:00: rank \#84 2026\-10\-01T15:00:51\.048047\+00:00: rank \#85 2026\-10\-01T16:00:53\.034654\+00:00: rank \#86 2026\-10\-01T16:36:37\.180071\+00:00: rank \#87 2026\-10\-01T17:00:23\.382263\+00:00: rank \#87 2026\-10\-01T18:01:44\.046011\+00:00: rank \#88 2026\-10\-01T19:01:47\.420738\+00:00: rank \#89 2026\-10\-01T20:01:39\.585083\+00:00: rank \#88 2026\-10\-01T21:01:45\.872085\+00:00: rank \#89 2026\-10\-01T22:01:37\.214899\+00:00: rank \#88 2026\-10\-01T23:04:11\.395288\+00:00: rank \#91 2026\-10\-02T08:02:26\.819307\+00:00: rank \#92 2026\-10\-02T09:02:51\.51649\+00:00: rank \#94 2026\-10\-02T10:02:12\.346099\+00:00: rank \#95 2026\-10\-02T11:02:22\.794376\+00:00: rank \#95 2026\-10\-02T12:01:14\.681578\+00:00: rank \#95 2026\-10\-02T13:02:29\.511591\+00:00: rank \#95 2026\-10\-02T14:01:40\.936738\+00:00: rank \#94 2026\-10\-02T15:02:30\.225476\+00:00: rank \#94 2026\-10\-02T16:02:47\.060695\+00:00: rank \#93 2026\-10\-02T17:02:31\.351493\+00:00: rank \#95 2026\-10\-02T18:01:22\.271899\+00:00: rank \#96 2026\-10\-02T19:01:23\.681417\+00:00: rank \#97 2026\-10\-02T20:01:33\.801055\+00:00: rank \#99 2026\-10\-02T21:02:55\.453316\+00:00: rank \#102 2026\-10\-02T22:01:50\.233139\+00:00: rank \#103 2026\-10\-02T23:01:20\.268516\+00:00: rank \#101 2026\-10\-03T08:01:30\.186188\+00:00: rank \#102 2026\-10\-03T09:01:57\.91467\+00:00: rank \#102 2026\-10\-03T10:00:55\.732001\+00:00: rank \#101 2026\-10\-03T11:01:22\.538596\+00:00: rank \#101 2026\-10\-03T12:01:08\.789384\+00:00: rank \#102 2026\-10\-03T13:00:35\.273559\+00:00: rank \#102 2026\-10\-03T14:00:48\.425606\+00:00: rank \#102 2026\-10\-03T15:01:04\.472963\+00:00: rank \#102 2026\-10\-03T16:01:56\.56175\+00:00: rank \#103 2026\-10\-03T17:01:36\.651036\+00:00: rank \#100 2026\-10\-03T18:01:58\.24207\+00:00: rank \#100 2026\-10\-03T19:01:19\.196589\+00:00: rank \#100 2026\-10\-03T20:00:57\.313759\+00:00: rank \#99 2026\-10\-03T21:01:04\.44439\+00:00: rank \#98 2026\-10\-03T21:20:01\.585706\+00:00: rank \#98 2026\-10\-03T22:00:15\.503126\+00:00: rank \#98 2026\-10\-03T23:00:55\.584832\+00:00: rank \#98 2026\-10\-04T08:01:00\.327459\+00:00: rank \#98 2026\-10\-04T09:00:59\.321023\+00:00: rank \#99 2026\-10\-04T10:00:42\.975336\+00:00: rank \#99 2026\-10\-04T11:01:03\.76577\+00:00: rank \#99 2026\-10\-04T12:01:45\.307985\+00:00: rank \#99 2026\-10\-04T13:00:49\.176267\+00:00: rank \#99 2026\-10\-04T14:00:56\.952543\+00:00: rank \#99 2026\-10\-04T15:00:54\.739637\+00:00: rank \#100 2026\-10\-04T16:00:51\.403054\+00:00: rank \#101 2026\-10\-04T17:00:43\.166713\+00:00: rank \#101 2026\-10\-04T18:00:46\.315535\+00:00: rank \#100 2026\-10\-04T19:00:54\.768841\+00:00: rank \#100 2026\-10\-04T20:00:51\.070442\+00:00: rank \#101 2026\-10\-04T21:00:28\.550532\+00:00: rank \#100 2026\-10\-04T22:01:36\.217651\+00:00: rank \#102 2026\-10\-04T23:01:09\.400767\+00:00: rank \#102 2026\-10\-05T08:01:51\.387003\+00:00: rank \#102 2026\-10\-05T09:01:40\.528475\+00:00: rank \#102 2026\-10\-05T10:00:56\.290772\+00:00: rank \#102 2026\-10\-05T11:01:05\.919518\+00:00: rank \#102 2026\-10\-05T12:00:56\.590881\+00:00: rank \#102 2026\-10\-05T13:01:05\.197156\+00:00: rank \#104 2026\-10\-05T14:01:04\.085718\+00:00: rank \#104 2026\-10\-05T15:00:50\.814571\+00:00: rank \#104 2026\-10\-05T16:02:09\.54139\+00:00: rank \#103 2026\-10\-05T17:00:35\.336076\+00:00: rank \#103 2026\-10\-05T18:00:20\.445491\+00:00: rank \#103 2026\-10\-05T19:02:43\.604975\+00:00: rank \#104 2026\-10\-05T20:01:09\.777459\+00:00: rank \#104 2026\-10\-05T21:01:47\.650671\+00:00: rank \#105 2026\-10\-05T22:02:59\.93099\+00:00: rank \#106 2026\-10\-06T08:02:11\.210914\+00:00: rank \#106 2026\-10\-06T08:02:14\.354697\+00:00: rank \#106 2026\-10\-06T09:01:21\.688765\+00:00: rank \#108 2026\-10\-06T10:00:58\.047773\+00:00: rank \#108 2026\-10\-06T11:00:41\.689929\+00:00: rank \#109 2026\-10\-06T12:00:41\.7705\+00:00: rank \#110 2026\-10\-06T13:01:24\.509519\+00:00: rank \#110 2026\-10\-06T14:00:26\.950817\+00:00: rank \#110 2026\-10\-06T15:00:49\.066414\+00:00: rank \#111 2026\-10\-06T16:00:58\.166747\+00:00: rank \#113 2026\-10\-06T17:01:17\.009511\+00:00: rank \#113 2026\-10\-06T18:01:43\.791429\+00:00: rank \#114 2026\-10\-06T19:00:46\.048609\+00:00: rank \#114 2026\-10\-06T20:01:14\.687294\+00:00: rank \#115 2026\-10\-06T21:02:13\.414382\+00:00: rank \#115 2026\-10\-06T22:00:40\.926271\+00:00: rank \#114 2026\-10\-06T23:03:26\.867445\+00:00: rank \#116 2026\-10\-07T08:01:41\.339804\+00:00: rank \#117 2026\-10\-07T09:02:38\.026986\+00:00: rank \#119 2026\-10\-07T10:01:35\.069531\+00:00: rank \#119 2026\-10\-07T11:02:07\.770084\+00:00: rank \#119 2026\-10\-07T12:01:14\.996697\+00:00: rank \#118 2026\-10\-07T13:00:47\.45912\+00:00: rank \#117 2026\-10\-07T14:01:18\.307739\+00:00: rank \#117 2026\-10\-07T15:01:57\.541362\+00:00: rank \#120 2026\-10\-07T16:01:28\.474022\+00:00: rank \#120 2026\-10\-07T17:01:15\.036908\+00:00: rank \#121 2026\-10\-07T18:00:53\.300079\+00:00: rank \#121 2026\-10\-07T19:01:49\.456657\+00:00: rank \#122 2026\-10\-07T20:01:59\.032487\+00:00: rank \#122 2026\-10\-07T21:05:36\.655392\+00:00: rank \#124 2026\-10\-07T22:02:27\.105241\+00:00: rank \#125 2026\-10\-07T23:01:17\.452894\+00:00: rank \#125 2026\-10\-08T08:03:55\.083808\+00:00: rank \#125 2026\-10\-08T09:03:58\.68898\+00:00: rank \#124 2026\-10\-08T10:04:05\.364235\+00:00: rank \#125 2026\-10\-08T11:02:04\.410307\+00:00: rank \#124 2026\-10\-08T12:02:36\.39618\+00:00: rank \#124 2026\-10\-08T13:03:09\.206401\+00:00: rank \#124 2026\-10\-08T14:03:30\.113384\+00:00: rank \#125 2026\-10\-08T15:02:54\.444758\+00:00: rank \#123 2026\-10\-08T16:03:13\.067807\+00:00: rank \#122 2026\-10\-08T17:03:47\.92465\+00:00: rank \#121 2026\-10\-08T18:01:50\.360431\+00:00: rank \#121 2026\-10\-08T19:02:55\.719274\+00:00: rank \#122 2026\-10\-08T20:01:36\.359022\+00:00: rank \#122 2026\-10\-08T21:03:34\.000719\+00:00: rank \#124 2026\-10\-08T22:01:13\.459732\+00:00: rank \#124 2026\-10\-08T23:01:25\.318392\+00:00: rank \#123 2026\-10\-09T08:01:53\.314021\+00:00: rank \#124 2026\-10\-09T09:02:59\.93426\+00:00: rank \#125 2026\-10\-09T10:02:22\.652527\+00:00: rank \#124 2026\-10\-09T11:02:04\.807282\+00:00: rank \#124 2026\-10\-09T12:02:16\.769295\+00:00: rank \#124 2026\-10\-09T13:02:30\.35034\+00:00: rank \#126 2026\-10\-09T14:01:34\.215134\+00:00: rank \#126 2026\-10\-09T15:01:27\.564392\+00:00: rank \#126 2026\-10\-09T16:02:06\.561056\+00:00: rank \#127 2026\-10\-09T17:03:06\.673648\+00:00: rank \#129 2026\-10\-09T18:01:02\.851052\+00:00: rank \#129 2026\-10\-09T19:01:11\.250291\+00:00: rank \#130 2026\-10\-09T20:01:32\.91907\+00:00: rank \#130 2026\-10\-09T21:02:08\.376144\+00:00: rank \#132 2026\-10\-09T22:01:19\.253075\+00:00: rank \#134 2026\-10\-09T23:01:31\.990912\+00:00: rank \#134 2026\-10\-10T08:01:00\.593376\+00:00: rank \#133 2026\-10\-10T09:01:45\.527975\+00:00: rank \#135 2026\-10\-10T10:00:53\.799353\+00:00: rank \#135 2026\-10\-10T11:00:35\.65851\+00:00: rank \#134 2026\-10\-10T12:01:11\.169647\+00:00: rank \#134 2026\-10\-10T13:00:54\.008377\+00:00: rank \#134 2026\-10\-10T14:01:43\.208766\+00:00: rank \#136 2026\-10\-10T15:01:24\.825188\+00:00: rank \#136 2026\-10\-10T16:01:10\.213328\+00:00: rank \#135 2026\-10\-10T17:00:48\.84189\+00:00: rank \#134 2026\-10\-10T18:00:57\.55132\+00:00: rank \#134 2026\-10\-10T19:00:52\.379463\+00:00: rank \#133 2026\-10\-10T20:00:21\.993912\+00:00: rank \#133 2026\-10\-10T21:01:10\.657738\+00:00: rank \#134 2026\-10\-10T22:00:50\.433569\+00:00: rank \#133 2026\-10\-10T23:01:00\.95027\+00:00: rank \#134 The discussion is largely skeptical of the tweet's implied degradation, mixing personal anecdotes of worse model behavior with demands for controlled benchmarks and alternative explanations such as capacity constraints, model transitions, or perception bias; no single cause is established\. ## The brief A tweet by Lon Lundgren claims that after Anthropic made Fable 5 permanently available in subscription plans, performance dropped noticeably\. The author says the model felt dumber and that five different measurement approaches showed August delivered dramatically fewer thinking tokens than July\. - Anthropic made Fable 5 permanently available in subscription plans\. - The author noticed a large drop in performance after that change\. - The model subjectively felt dumber, which the author could not initially explain\. - Measured five different ways, August reportedly delivered dramatically fewer thinking tokens than July\. ## Discussion themes ### Anecdotal degradation extends beyond Fable 5 Several commenters report similar perceived declines across models\. alexjplant recalls Anthropic denying performance throttling while seeing Opus peaks and troughs; mlmonkey describes frontier models becoming less capable over weeks; jotato says gpt\-5\.6\-luna needed more explicit instructions; Waterluvian reports Fable 5 duplicating instead of deleting and mixing up predicate names\. Sources: [Comment 49789574]() · [Comment 49789794]() · [Comment 49789952]() · [Comment 49790055]() ### Methodological objection: non\-determinism and perception cloudking asks how to create repeatable tests in a non\-deterministic system where the same prompt yields different answers\. JMKH42 argues human perception is poor at detecting trends and calls for rigorously tested benchmarks with enough trials to know standard deviation, noting that memory or project growth can pollute results\. Sources: [Comment 49789843]() · [Comment 49790792]() ### Alternative explanations: capacity and model transitions prodigycorp says a new Fable and Opus 5\.5 release is pending and Anthropic is reallocating resources, with Opus 5\.5 served under Opus 5\. briffle attributes the issue to cloud vendors running at full capacity and training newer models rather than malice\. SequoiaHope asks what mechanism would degrade capability instead of returning a resource error\. Sources: [Comment 49790119]() · [Comment 49790899]() · [Comment 49790476]() ### Allegations of intentional degradation or scam\-like behavior r2\-129 describes a pattern of models being quantized or given worse inference after launch\. talon8635 speculates about slowly dumbing down a model so a later marginal release looks better\. AmazingTurtle agrees and is shifting to open\-weight models for stable quality\. fidotron says Anthropic is not a good actor and questions blind trust in it\. Sources: [Comment 49789703]() · [Comment 49790717]() · [Comment 49790762]() · [Comment 49790668]() ### Accountability and benchmarking proposals saejox wanted regular private benchmarks of hosted models but sees no monetization path and expects labs to game such a service\. jesse\_dot\_id invokes the Office of Weights and Measures as a regulatory analogy for inconsistent products\. physicallyIllfr compares model use to slot machines and says the UX encourages repeated spending\. Sources: [Comment 49790330]() · [Comment 49790471]() · [Comment 49790431]() ### Technical details raised espeed notes that Claude Code's prompt cache expires after 1 hour\. zeroonetwothree doubts a 2x DGX Spark cluster would pay back before the hardware dies\. Sources: [Comment 49790830]() · [Comment 49790985]() ## Sources & coverage AI-generated summary · 2026\-09\-21T20:01:24\.286025\+00:00 Based on 21 of 21 usable stored comments, selected by depth and branch activity. This is a sample of the discussion. Article text may also be shortened. Generated using deepseek\-ai/DeepSeek\-V4\.1\-Flash. Check the linked sources for full context.