My agent is working and giving answers, but the problem i have is it hit the 16k token limit (which is configured that way) I've tried /reset /clear, even deleting manually the files that contains the context but I'm still stuck at 16k on the graphic interface in openclaw, idk what to do, do someone have a solution? I'm not a developer just a mere technician.
#Hitting the 16k context window
1 messages · Page 1 of 1 (latest)
First: please paste your openclaw --version — version matters a lot here.
Important distinction: /reset / /new clears the session history, but it does not increase the model’s configured 16k context window. Also, OpenClaw’s base prompt/tools/workspace files count toward that 16k, so a 16k model can feel “full” even after a reset.
Try this exact flow:
- In that same UI chat, send only:
/new - Then send:
/status /context list - If
/context listis already huge on a fresh session, the fix is to reduce baseline context or use a bigger-context model:- trim big
AGENTS.md/TOOLS.md/ memory files
- trim big