deepseek-v4-flash does not have max output tokens as 8192 according to their own API. This needs another look!

https://github.com/BerriAI/litellm/blob/{"message":"API rate limit exceeded for 114.143.116.18. (But here's the good news: Authenticated requests get a higher rate limit. Check out the documentation for more details.)","documentation_url":"https://docs.github.com/rest/overview/resources-in-the-rest-api#rate-limiting"} /model_prices_and_context_window.json#L43811-L43835

We also need to update [model_prices_and_context_window_backup.json](https://github.com/BerriAI/litellm/blob/{"message":"API rate limit exceeded for 114.143.116.18. (But here's the good news: Authenticated requests get a higher rate limit. Check out the documentation for more details.)","documentation_url":"https://docs.github.com/rest/overview/resources-in-the-rest-api#rate-limiting"} /litellm/model_prices_and_context_window_backup.json).