Hi all, I’m using Vercel AI Gateway with DeepSeek v3.2 for long “reasoning/thinking” tasks and I’m consistently hitting a hard cutoff:
Non-streaming requests get terminated at ~340s
Streaming runs longer but still gets cut off at ~790s (stream disconnects mid-output)
Has anyone seen this? What’s the usual way to fix or work around it (gateway timeout settings, async job + polling, splitting the task, etc.)?
Thanks!
Hi , welcome to the Vercel Community!
Is it because your Vercel Functions are timing out? Or is it happening locally as well?
Please share your public repo or a minimal reproducible example. That will let us all work together from the same code to figure out what’s going wrong.
Hi , just checking in on your timeout issue with DeepSeek v3.2 on Vercel AI Gateway. Have you found a solution or still need assistance? Let me know if you need more details.