Not getting any response...

[Ruko2010]

I have set up an LLM Proxy with LiteLLM to manage all the AI Access, which is working fine.
I created a Virtual Key, just for this setup.
The API and Models are fine if I call them from other programms.

My uv config file looks like this:

OPENAI_API_KEY="sk-W..."
PREFERRED_PROVIDER="openai"
OPENAI_BASE_URL="http://XXX.YYY.com:4000"
BIG_MODEL="spark/glm-4.7-flash"
SMALL_MODEL="spark/glm-4.7-flash"

Now, If I start Claude Code with this settings:

export ANTHROPIC_AUTH_TOKEN="ollma"
export ANTHROPIC_API_KEY=""
export ANTHROPIC_BASE_URL=http://0.0.0.0:8082
export CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1
claude

I see, that uv has startet and taking the Requests from Claude:

POST /v1/messages ✓ 200 OK
claude-haiku-4-5-20251001 → glm-4.7-flash 0 tools 2 messages
POST /v1/messages ✓ 200 OK
claude-sonnet-4-6 → glm-4.7-flash 20 tools 2 messages
POST /v1/messages ✓ 200 OK
claude-haiku-4-5-20251001 → glm-4.7-flash 0 tools 2 messages
POST /v1/messages ✓ 200 OK
claude-sonnet-4-6 → glm-4.7-flash 20 tools 2 messages

If I type something in Claude, like "hey", I see two new Lines, which indicate, the uv is getting the request from claude

POST /v1/messages ✓ 200 OK
claude-haiku-4-5-20251001 → glm-4.7-flash 0 tools 2 messages
POST /v1/messages ✓ 200 OK
claude-sonnet-4-6 → glm-4.7-flash 20 tools 2 messages

But, I never get an output in claude:

In LiteLLM, I see that the Request is getting there, and it is successful:

Is there a way I can debug this? I have no Idea where the error could be.