Deepseek V4 Pro Model Usage Issue, "Error: 500 Network connection lost."

Summary

I was about to use Deepseek V4 Pro, and when I sent my prompt it gave out the 500 error, re-tried again with another prompt but it sent the same error.

Trace IDS: 020171296a322005cec9aa990316a69a and 7955043702a563824f7861930700848d

Expected Behavior

It shoud make / generate the code planned in the prompt, but it gave that error exception

Actual Behavior

I sent the prompt, It took a long time to parse it (and it didn't generate anything), I was waiting the prompt and I've found weird that the token consumption was 0 and did a ctrl + o., and the context of what the model was thinking wasn't appearing which meant that my petition didn't go anywhere.

Steps to reproduce the issue

1. Prompt (Whatever you want)
2. Just wait for the "500 Error Exception"

Command Code Version

Unknown

Operating System

Windows

Terminal/IDE

Orca

Shell

cmd.exe

Additional context

I'm using 1.4.4 Version

Reproduce the failure in Orca on Windows using Command Code 1.4.4 and a simple prompt, then inspect the 500 response and the two supplied trace IDs. Compare the behavior with Deepseek V4 Pro and check whether the request produces tokens or context; done means the prompt completes instead of returning “500 Network connection lost.”