Frontier model capability climbs in stacked S-curves: Fable 5.2 hits 173.4, GPT-6 Astra 169.0
scaling01 · x · 2026-11-16
scaling01 shares an era-based analysis of frontier capability (ECI index) for OpenAI and Anthropic models:
- Anthropic: three S-curve eras — Sonnet 3.5→4 (ceiling 142.0), Opus 4→4.8 (ceiling 159.9), Fable 5→now (ceiling 176.9, latest Fable 5.2 at 173.4).
- OpenAI: GPT-4o→5.1 (ceiling 152.0), GPT-5.2→5.4 Sol (ceiling 164.2), GPT-6 Astra→now (ceiling 178.8, double-S shape; Astra at 169.0).
- A companion scatter of no-CoT reasoning vs knowledge shows newest-generation models (Fable 5.1, GPT-6 Astra) far above the all-model trend line.
The author proposes plotting the reasoning era, agentic-RL-maxxing era, and loopmaxxer/multi-agent era as separate S-curves, notes it's confusing to include non-frontier models like GPT-5.1, and quips that Anthropic probably has no loopmaxxer-era model to name.
Related event: New hypothesis: AI capability growth comes in chained S-curves(2 posts)→
More from Models
- Rumor: Google Prepping Strongest Gemini Model Yet, Codenamed Krypton — aziz4ai · 2026-11-16
- Blender watch test: DeepSeek does it for $0.80 vs Claude Fable's $119 — HankYeomans · 2026-11-16
- Users say Gemini turned colder after Nov 12 update, want old version back — Sean8277 · 2026-11-16
- Scaling曲线新假说:三段S曲线接力,外推天花板或达178+ — scaling01 · 2026-11-16
- Grok rewrite reportedly degrades at understanding writing intent — xiaosun86 · 2026-11-16
- Google's Gemini shifts to element-themed model names: Carbon, Argon, Hydrogen, Krypton — Dense_Primary629 · 2026-11-16