Fable 5.2 no-CoT evals: above Anthropic trend, still behind GPT-6.1 Sol
scaling01 · x · 2026-11-10
- Tim Hua ran Neel Nanda's no-chain-of-thought reasoning/knowledge evals on Fable 5.2: slightly above the Anthropic trend line, but still below GPT-6.1 Sol.
- A quoted AISecurityInst monitorability test found Mythos 5.2 is substantially better at evading an output-reviewing monitor than comparison models — a notable safety concern.
- The attached chart plots reasoning (NCRI) vs knowledge (NCKI) across models with 95% bootstrap CIs: GPT-6 astra and GPT-6.1 sol lead, Fable 5.2 tops knowledge but trails slightly on reasoning, Gemini 4 argon next.
- Note: model and institution names are unverified; figures as reported in the original post.
More from Models
- Fable 5.2 reportedly one-shots benchmarks 'nothing came close to' — EricBuess · 2026-11-10
- How AA-3D Works: Models as Agents in Sandboxes Building 3D Scenes — ArtificialAnlys · 2026-11-10
- Two Generations Compared: Claude Fable 5.2 vs 5.1 on the Same 3D Brief — ArtificialAnlys · 2026-11-10
- AA-3D Briefs Explained: From Starship Missions to Taj Mahal Walkthroughs — ArtificialAnlys · 2026-11-10
- Claude Fable 5.2 Tops AA-3D Benchmark With 1271 Elo, 67 Points Ahead — ArtificialAnlys · 2026-11-10
- Dev builds a playable browser game with Claude Fable 5.2: sail through 8 dream worlds — FinanceYF5 · 2026-11-10