Problem
What problem does this solve?
众所周知，qoder的上下文压缩后会导致丢失很多信息，并导致会话质量急剧下降。现在qoder触发压缩的上下文长度是200k以下，但一些模型已经有了较多的上下文长度。比如deepseek v4 (flash/pro)。如果能够较晚进行压缩，能够很好地提升很多情况下agent的会话质量。
Solution
How should it work?
希望对某些模型，触发上下文压缩的上下文长度阈值更大一些
Use Case
When would you use this?
一些长上下文的会话。用于优化会话质量。
Priority
-
High - Blocking issue -
Medium - Important improvement -
Low - Nice to have
Additional Info
(Optional: screenshots, examples, links)