AnalysisAI ModelsNovember 18, 2026

Test finds Claude models trained on protein sequences at scale

Read original source →xinmingtu.cn

On 668 little-studied proteins, Claude Fable 5.2 predicted the next amino acid 54% of the time versus 9% on shuffled copies; Opus 5.5 matched ProtGPT3-10B. GPT, Gemini and seven open-weight models showed no such signal.

1 source

More stories today

Open the live feed