Fable 5.1 more than doubles score on agentic scientific workflows, 24.7% to 52.6%

haider1 · x · 2026-09-02

haider1 reports that Fable 5.1 jumped from 24.7% to 52.6% on agentic scientific workflows — more than 2x better than its predecessor. He ties this to RL prioritization, noting rumors that OpenAI's unreleased models are exceptionally strong at math, suggesting labs can push fields one by one by choosing what to optimize.

Original post →

More from Models

Models channel →