Auto model execution failure due to reasoning level

[Daerdor]

Describe the bug

I was using Claude Sonnet 4.6 using 'High Reasoning'

I switched the model to 'Auto' and it initially kept using Claude Sonnet 4.6 so there was no issue.
But then it switched to Claude Haiku 4.5 and produced the following error:

"Execution failed: CAPIError: 400 reasoning_effort "high" was provided, but model claude-haiku-4.5 does not support reasoning effort (Request ID: B2B7:2BE0B7:3ADE121:42B996B:69E62C35)"

Auto mode should automatically disable the reasoning level in case a chosen model does not support this.

Affected version

1.0.33

Steps to reproduce the behavior

See above:

1. Use a model with high reasoning
2. Switch to auto model
3. When auto automatically chooses a model with no reasoning the error appears

Expected behavior

Auto mode should automatically disable the reasoning level in case a chosen model does not support this.
The reasoning level should be stored so that, when another model is chosen that does support reasoning, this level is kept.

Additional context

No response

[devm33]

Hi @Daerdor thanks for raising this. It should be fixed in the next release. Let us know if you see any further issues.