Spring-ai Add Support for Cached Tokens in OpenAI Usage Model

Hey team,

Could you add support for cached_tokens in the Usage model for OpenAI responses?

The new response structure from OpenAI looks like this: "usage": { "prompt_tokens": 2006, "completion_tokens": 300, "total_tokens": 2306, "prompt_tokens_details": { "cached_tokens": 1920 }, "completion_tokens_details": { "reasoning_tokens": 0 } }

For more details, you can refer to the official OpenAI documentation on prompt caching: OpenAI Prompt Caching Guide

This would be a valuable addition to enhance the accuracy of token usage tracking.

Thanks!