Summary

What went wrong?
Tried to ask via Chat (original model is Qwen) to change to local Ollama's deepseek-r1:latest, which is already pulled and verified.

Steps to reproduce

1. Installed Ollama and run ollama pull deepseek-r1:latest
2. Ask in chat to change the default model to DeepSeek (also tried Llama3.3, same problem)
3. The model cannot be changed. In CLI, I run openclaw model list, it reports:

% openclaw models list        

🦞 OpenClaw 2026.1.29 (a5b4d22) — One CLI to rule them all, and one more restart because you changed the port.

Model                                      Input      Ctx      Local Auth  Tags
ollama/deepseek-r1:latest                  -          -        -     -     default,missing
qwen-portal/vision-model                   text+image 125k     no    yes   fallback#1,configured
qwen-portal/coder-model                    text       125k     no    yes   configured,alias:qwen

It said the Ollama Deepseek is missing.

Expected behavior

What did you expect to happen?
Expect the model changed and restart the gateway, then everything is fine.

Actual behavior

What actually happened?
The chat stops responding afterwards.

Environment

• Clawdbot version: 2026.1.29
• OS: macOS Tahoe 26.2
• Install method (pnpm/npx/docker/etc): npm install -g openclaw@latest

Logs or screenshots

Paste relevant logs or add screenshots (redact secrets).

The output of ollama list is:

% ollama list
NAME                  ID              SIZE      MODIFIED       
deepseek-r1:latest    6995872bfe4c    5.2 GB    14 minutes ago