Claude Fable 5.2 debuts with 74.6% on Terminal-Bench-Science, topping rival models
BenBlaiszik · x · 2026-11-10
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2, aimed at the most daunting technical challenges with gains in coding, reasoning, and vision. A shared benchmark table shows Fable 5.2 scoring 74.6% on Terminal-Bench-Science 0.1, plus leading results like 69.9% FrontierSWEv2, 85.1% de novo binder protein design (Mythos 5.2), 89.3% OSWorld 2.1 computer use, 64.9% RoboDojo Sim robotics, and 74.2% Humanity's Last Exam with tools — beating GPT-6 Astra, Gemini 4 Carbon, and prior versions on most axes. The Terminal-Bench team is preparing a harder next release.
Related event: Anthropic Launches Claude Fable 5.2 and Mythos 5.2(40 posts)→
More from Models
- Anthropic gives loyal Claude Code subscribers up to 3 Fable resets based on tenure — threepointone · 2026-11-10
- Users ask if Claude Opus 5.5 has already been quietly nerfed — eigenron · 2026-11-10
- Ant Group's Robbyant open-sources LingBot Touch, a JEPA-based tactile foundation model — AdinaYakup · 2026-11-10
- Fable 5.2 not yet showing up in Claude Code for some users — DavidmComfort · 2026-11-10
- Claude Fable 5.2 designs 3 working electronic devices from one-paragraph prompts — Dr_Singularity · 2026-11-10
- Salesforce's Interactive Visual Gym: best VLM finishes only 53.4% of 315 closed-loop tasks — jcniebles · 2026-11-10