Hello everyone, I’m currently using Windows with the latest version of Cursor in a WSL2 environment.
When using the new ‘Deepseek R1’ model, I encounter issues like incomplete responses. I’m not sure if the limitation is caused by Cursor or if it’s an issue inherent to Deepseek itself. If it’s the latter, it would be great to implement a feature that allows responses to continue seamlessly.
I think it may be a time out issue. I have seen this before and the response got cut off during the thinking phase
Hey, thanks for the bug report. The DeepSeek R1 model is currently in experimental mode and may produce errors like this. I hope it will be fixed in the future.
Thanks for the reports on this!
We’re aware that a timeout on our end is causing R1 requests to be canceled before it is finished, but we are hoping to increase this in the next few days or so, to allow R1 the time to complete its output.
We are also working on improving the speed of both v3 and R1 in Composer, but this could take longer to get our infrastructure where it needs to be to do so.
DeepSeek R1 stops in the middle of thinking for me… Sad that we cannot utilize such great model yet…
One workaround I found is this:
When R1 stops in the middle of thinking, I send a message:
You haven’t finished thinking! continue please
and it works! The model continues thinking then producing the final output.
Hope it helps while the Cursor team fix these issues as mentioned by @danperks
