Nous Research shares Fable 5.2 benchmark table spanning 17 tasks against Opus 5.5 and GPT-6
dean_rie · x · 2026-11-10
Nous Research says Fable 5.2 is now live in Hermes Agent and posted a 17-benchmark scorecard comparing it against Fable 5.1, Claude Opus 5.5, GPT-6 Astra and Gemini 4 Carbon. Claimed highlights: 69.9% on FrontierSWE v2 agentic coding, 74.6% on Terminal-Bench-Science, 89.6% on CVE Rediscovery, 89.3% on OSWorld 2.1 computer use, and 41.3% on the Remote Labor Index. Note that several compared models are unreleased and the scores are unverified.