### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### The Feature

As per https://docs.x.ai/developers/advanced-api-usage/prompt-caching, Grok recommends setting `x-grok-conv-id` request parameter to maximize odds of cache hits. I believe they added this param in March 2026.

Suggestion for LiteLLM: 1) Document and explicitly support passing this named parameter for X.Ai provider, not requiring extra_headers etc 2) Also support passing this parameter to OpenRouter when x.ai/grok models are called via OpenRouter.

It's not obvious whether OR would already forward the header (maybe not yet) but worth it setting it regardless, OpenRouter might improve in the future.

### Motivation, pitch

Improve odds for cache hits for X-Ai Grok following their official API documentation.

### What part of LiteLLM is this about?

Proxy

### LiteLLM is hiring a founding backend engineer, are you interested in joining us and shipping to all our users?

No

### Twitter / LinkedIn details

_No response_