New Features

• Added an option reuse_client to openai/azure to help with async timeouts. Set to False to see improvements (#9301)
• Added support for vLLM llm (#9257)
• Add support for python 3.12 (#9304)
• Support for claude-2.1 model name (#9275)

Bug Fixes / Nits

• Fix embedding format for bedrock cohere embeddings (#9265)
• Use delete_kwargs for filtering in weaviate vector store (#9300)
• Fixed automatic qdrant client construction (#9267)