Worked on improving the stability and correctness of OpenAI model filtering within the LibreChat repository. Focused on a targeted bug fix that updated the ModelService regex logic to include model names beginning with 'o1-', ensuring more accurate filtering of OpenAI models in API responses. Utilized JavaScript and regular expressions to refine the API’s selection process, directly addressing issues with incorrect model selection by clients. This change enhanced the reliability of LibreChat’s OpenAI integration, resulting in more predictable behavior in production. The work contributed to increased user trust and reduced support overhead by making model filtering more robust and maintainable.