Employing vLLM native RL APIs

**Is your feature request related to a problem? Please describe.**

We need to employ the [Improved Pause/Resume Support for Asynchronous RL](https://vllm.ai/blog/2026-05-28-native-rl-apis) APIs for pause/resume and recompute the KV cache when our vLLM version becomes higher to use these APIs.

**Describe the solution you'd like**
A clear and concise description of what you want to happen.

**Describe alternatives you've considered**
A clear and concise description of any alternative solutions or features you've considered.

**Additional context**
Add any other context or screenshots about the feature request here.