Genering tests with new models of OpenAI (gpt-5.3-codex, gpt-5.2-codex,..) leads to 404

### Involved Module
- [x] LLM

### Describe the bug
Genering tests with new models of OpenAI (gpt-5.3-codex, gpt-5.2-codex,..) leads to 404: Request failed with status code 404

### To Reproduce
Steps to reproduce the behavior:
1. Open class under test
2. Right-click on the class -> choose TestSpark
3. Choose LLM - choose model gpt-5.3-codex - OK to generate tests
4. See error: Request failed with status code 404
Note: with old models such as gpt-5: still work ok

### Expected behavior
generate tests successful

### Screenshots
If applicable, add screenshots to help explain your problem.

### Additional context
Add any other context about the problem here.

Start by reproducing the TestSpark workflow from an IDE class using gpt-5.3-codex or gpt-5.2-codex, then inspect the LLM integration and model request handling. The fix is complete when these newer models generate tests successfully without a 404 while the existing gpt-5 flow continues to work.

Written by the indexing model from the issue text.