400 {"error":"\"qwen2.5:0.5b\" does not support thinking"}

TL;DR

The issue might be resolved by checking the compatibility of the model "qwen2.5:0.5b" with the "thinking" feature in OpenClaw.

Guidance

• Verify that the model "qwen2.5:0.5b" supports the "thinking" feature by checking the model's documentation or configuration.
• Check the OpenClaw configuration to ensure that the "thinking" feature is enabled for the model "qwen2.5:0.5b".
• Test the model with a different prompt that does not require the "thinking" feature to see if the issue is specific to this feature.
• Investigate why direct Ollama calls to /api/chat with the same model work, including with "think": false, to identify potential differences in configuration or implementation.

Example

No code snippet is provided as the issue does not contain sufficient information to create a relevant example.

Notes

The issue lacks detailed information about the model configuration, OpenClaw setup, and logs, which makes it difficult to provide a more specific solution. Further investigation and debugging are required to identify the root cause of the issue.

Recommendation

Apply workaround: Try using a different model that is known to support the "thinking" feature or disable the "thinking" feature for the current model to see if it resolves the issue. This is because the error message specifically mentions that the model does not support thinking, suggesting a potential compatibility issue.