Error when i use deepseek-v4-flash

Hey, this is a known issue with BYOK and DeepSeek V4 in thinking mode: our proxy isn’t passing reasoning_content back in multi-turn requests with tool calls, and the DeepSeek API rejects them with invalid_request_error. I reported the issue internally, no ETA for a fix yet.

A workaround was already shared earlier in this thread: the proxy by @yxlao that reinjects reasoning_content: GitHub - yxlao/deepseek-cursor-proxy: Fixes reasoning_content issue of DeepSeek V4, supprting deepseek-v4-pro and deepseek-v4-flash · GitHub

Main discussion and extra context: Compatibility with DeepSeek model's design to return reasoning content after tool calls?