#ChatGPT 5.2 xhigh reasoning freezes endlessly

4 messages · Page 1 of 1 (latest)

icy ether
#

When using ChatGPT 5.2 xhigh reasoning, Warp constantly freezes for like 10-30 seconds or longer. There is some process that should be async to prevent the UI from freezing

soft kettleBOT
#

Hi @icy ether. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

#

There’s not yet a specific, documented fix for freezes specifically with “ChatGPT 5.2 xhigh reasoning” in the knowledge sources, but there are several closely related, known issues where Warp’s AI / planning models or Codex usage cause the UI to appear frozen or unresponsive.

Relevant patterns from existing reports:

  • Planning model freezes (GPT‑5): When using GPT‑5 as the planning model, the “planning step” can get stuck and the UI looks frozen until the user cancels with Ctrl‑C. [planning freeze]
  • Codex / AI long‑use freezes: Users report Warp freezing when using Codex or AI for a long time; this is confirmed to be Warp‑specific. [codex freeze]
  • Chat / agent “freezing”:
    • Chat UI sometimes stops accepting input and appears frozen. [chat freezing]
    • For long‑running commands, the agent waits for the command to finish and the UI doesn’t clearly show progress, which can feel like a freeze. [chat comments]
  • General UI freezes on Linux / GPU issues: Some users see Warp freezing or even system‑level stutters, sometimes with logs like The current device is lost. Reason: Unknown. Message: Unexpected error variant (driver implementation is at fault). [device lost warn; linux freeze; UI freeze]

From these, it’s clear Warp currently has multiple open issues where AI‑related work (planning, Codex, chat) or GPU/graphics problems can block or stall the UI instead of being fully asynchronous.

#

Because the sources don’t mention “ChatGPT 5.2 xhigh reasoning” by name, I can’t confirm that your exact model choice is a known, tracked bug. But your symptoms (10–30s+ UI freezes while an advanced reasoning model runs) are very similar to the GPT‑5 planning and Codex freeze reports above.

What you can do right now

Since there’s no documented workaround specific to your model in the sources, the best supported path is:

  1. File a bug with logs and model details

    • Include that it happens with “ChatGPT 5.2 xhigh reasoning,” how long the UI is blocked, and whether Ctrl‑C restores control.
    • Attach Warp logs as suggested in other issues. [planning freeze]

  2. Mention that it’s UI‑blocking (non‑async) work

    • Explicitly note that the UI thread appears blocked for 10–30 seconds while the model is reasoning, similar to the planning‑step and Codex freezes.

  3. If possible, test with another model

    • In the GPT‑5 planning issue, switching models or restarting Warp temporarily restored functionality. [planning freeze]
    • If a different model doesn’t freeze, include that comparison in your report.

Because the current knowledge sources don’t describe an existing fix or configuration to force this work fully async, I can’t provide a concrete setting or patch to resolve it today. The behavior you’re seeing is consistent with other open AI‑related freeze bugs that Warp is still working on.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon: