llama.cpp Adds RANK Pooling Batch Splitting for Causal LLM Rerankers (e.g., Qwen3)

Core summary

One sentence to understand this update

The latest llama.cpp server update introduces RANK pooling batch splitting to improve efficiency for causal LLM rerankers like Qwen3 and Qwen3-VL.

Impact & opportunity

What this could mean

Builders can now leverage more efficient reranking with causal LLMs on llama.cpp, potentially improving performance for retrieval-augmented generation.