ByteBrief
We're a portrait publication through and through. Turn your phone back and your briefing picks up right where you left it.
(We tried widescreen once. It wasn't us.)
Researchers introduce Context Language Models, which treat context as a file the model updates itself. Zero-shot CLMs beat state-of-the-art context management by 11.4% on BrowseComp-Plus with 21.5% fewer FLOPs, and 5% higher on 12-hour EdgeBench with 59% fewer FLOPs. Online reinforcement learning improves Qwen3.5-9B by 47.6% on BrowseComp-Plus using 12% fewer FLOPs.
Tap to vote and see what everyone thinks.
Summary by ByteBrief