[Bug] Thinking tags don't get formatted correctly
-
I'm using DeepSeek R1
-
The thinking tags of DeepSeek R1 get streamed as a normal response instead of being in a special collapsible thing like in most apps
Yeah, it's also an issue with QwQ model, particularly when Void tries to update the file, code is mixed with thinking output, which breaks the app
yeah and the output can mostly be long and annoying, having like a field like the "Continue" or "Roo Code" to hide the thinking output and show when needed would be great
You sure you're using 1.2.4? What DeepSeek provider are you using?
i'm using custom provider, using chutes.ai
I am using Deepseek R1 32B and the
I am using Deepseek R1 32B and the and tags do not get formatted correctly!
yeah that's the same problem with me