---
title: "Another deepseek v4.1 flash issue, model started to refuse work around 700k token line, saying..."
date: 2026-09-12
canonical: https://solmaz.io/x/2098598729687470107/
x_url: https://x.com/onusoz/status/2098598729687470107
license: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
---

Another deepseek v4.1 flash issue, model started to refuse work around 700k token line, saying "context is gone", and it has no context repeatedly

deepseek advertises 1m context, but the usable context with this model is probably still around 200-300k like other models. I've set that as max before compaction

*Quotes a post by @onusoz (https://x.com/onusoz/status/2098453692546683179); its text is omitted here because it is not covered by this site's license.*