{"slug": "fable-5-median-thinking-declined-in-august", "title": "Fable 5 – Median thinking declined in August", "summary": "Anthropic's Fable 5 model delivered dramatically fewer thinking tokens in August than in July after the company made the model permanently available in subscription plans, according to measurements by Lon Lundgren. Lundgren reported that reasoning declined over the entire period and fluctuated across multi-day episodes, with most invocations receiving little to no thinking tokens even at xhigh or max effort levels, and longer thinking runs almost never reaching published benchmark levels. Lundgren advised users to \"Ask about the inference regime you were served\" rather than assuming the model was nerfed.", "body_md": "Lon Lundgren on X: \"After Anthropic made Fable 5 permanently available in subscription plans, I noticed a large drop in performance. The model felt dumber, and I couldn't explain why.\nMeasured five different ways, August delivered dramatically fewer thinking tokens than July.\" / X\n\nLon Lundgren on X: \"After Anthropic made Fable 5 permanently available in subscription plans, I noticed a large drop in performance. The model felt dumber, and I couldn't explain why.\nMeasured five different ways, August delivered dramatically fewer thinking tokens than July.\"\n\nAfter Anthropic made Fable 5 permanently available in subscription plans, I noticed a large drop in performance. The model felt dumber, and I couldn't explain why.\nMeasured five different ways, August delivered dramatically fewer thinking tokens than July.\n\nAfter Anthropic made Fable 5 permanently available in subscription plans, I noticed a large drop in performance. The model felt dumber, and I couldn't explain why.\nMeasured five different ways, August delivered dramatically fewer thinking tokens than July.\n\nThis wasn't a one-time drop. Reasoning fell over the entire period, and fluctuated across multi-day episodes.\nSome of these fluctuations aligned with specific product announcements and releases.\nI started to see how the model could feel great one day, and terrible the next.\n\nI was consistently using an xhigh or max effot level, but when I looked deeper, I found that most invocations to the model were receiving little to no thinking tokens at all.\nAnd when longer thinking runs did happen, they almost never reached published benchmark levels.\n\nThis unfolded across a six week data capture and analysis odyssey, and led to a number of surprising findings.\nNext time the model feels dumber, don't ask if the model was \"nerfed\".\nAsk about the inference regime you were served, instead.\n\nBuddy its so simple. If one person used fable. It'll blow your socks off. Divided by millions It starts to suck. Models get dumber as usage goes up until they make another datacenter. Then on and on forever", "url": "https://wpnews.pro/news/fable-5-median-thinking-declined-in-august", "canonical_source": "https://twitter.com/Lon/status/2101793422487204027", "published_at": "2026-09-21 16:13:57+00:00", "updated_at": "2026-09-21 16:25:04.086160+00:00", "lang": "en", "topics": ["large-language-models", "ai-products", "ai-infrastructure"], "entities": ["Anthropic", "Fable 5", "Lon Lundgren"], "alternates": {"html": "https://wpnews.pro/news/fable-5-median-thinking-declined-in-august", "markdown": "https://wpnews.pro/news/fable-5-median-thinking-declined-in-august.md", "text": "https://wpnews.pro/news/fable-5-median-thinking-declined-in-august.txt", "jsonld": "https://wpnews.pro/news/fable-5-median-thinking-declined-in-august.jsonld"}}