# BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference > Source: > Published: 2026-09-09 05:30:22+00:00 Large Reasoning Models (LRMs) achieve superior problem-solving through extended Chain-of-Thought (CoT) generation, but the resulting key-value (KV) cache grows linearly with sequence length and creates severe memory bottlenecks, often exceeding GPU capacity for long reasoning traces. Existing KV cac