YANchor-4B: Effective Long-Horizon Reasoning in O(N) Time with O(1) Memory

arXiv:2610.10118v1 Announce Type: cross Abstract: Long-horizon reasoning demands access to earlier information at a manageable generation cost. Full-history attention incurs growing storage and computation, while recurrent compression can lose precise details. Therefore, we present YANchor-4B, a general-purpose recurrent model that preserves crucial memory as ANchors for retrieval during subsequent reasoning. Beyond O(N)-time generation and O(1) memory, YANchor enables effective long-horizon reasoning through its multidimensional memory mechanism. For example, on challenging math problems, it achieves 82.93% mean pass@1 on AIME 2024--2026 and 63.64% on HMMT, substantially outperforming linear-time, constant-state counterparts, including larger models. It also delivers several-fold higher batched long-generation throughput than Transformer and hybrid baselines on H100. Furthermore, evaluations across dozens of benchmarks demonstrate YANchor's superiority in general-purpose capabilities.

Related

• OScaR: The Occam's Razor for Extreme KV Cache Quantization in LLMs and Beyond
• BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference
• KVDrive: A Holistic Multi-Tier KV Cache Management System for Long-Context LLM Inference
• SparDA: Sparse Decoupled Attention for Efficient Long-Context LLM Inference

Source: arXiv cs.CL | 2026-10-08