Long-context models still waste tokens by copying the prompt into reasoning

eyishazyer · x · 2026-07-23

A post about a paper argues that long-context reasoning still has a hidden failure mode: even strong models copy large chunks of the prompt into their own chain-of-thought, wasting tokens and hurting accuracy.

Related event: Study Reveals Long-Context Models Waste Tokens Copying Text(2 posts)→

Original post →

More from Research

Research channel →