sync

[Matvezy]

What does this PR do?

Add selection ob base models fro inference_models-only aliases and drop thinking mode for 0.8 Qwen3.5 size due to token repetition.

Type of Change

• Bug fix (non-breaking change that fixes an issue)

Testing

• I have tested this change locally
• I have added/updated tests for this change

Test details:
Tested locally

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code where necessary, particularly in hard-to-understand areas
• My changes generate no new warnings or errors
• I have updated the documentation accordingly (if applicable)

Additional Context

No