# ARCQuant: Boosting NVFP4 Quantization with Augmented Residual Channels for LLMs Source: [ARCQuant: Boosting NVFP4 Quantization with Augmented Residual Channels for LLMs]() Collection index: [llms.txt]() Data updated: 2026-07-08T05:00:13.057Z Authors: Haoqian Meng, Yilun Luo, Yafei Zhao, Wenyuan Liu, Peng Zhang, Xindian Ma Published: 2026-01-12T12:27:22.000Z ## Abstract The emergence of fine-grained numerical formats like NVFP4 presents new opportunities for efficient Large Language Model (LLM) inference. However, it is difficult to adapt existing Post-Training Quantization (PTQ) strategies to these formats: rotation-based methods compromise fine-grained block isolation; smoothing techniques struggle with significant 4-bit quantization errors; and mixed-precision approaches often conflict with hardware constraints on unified-precision computation. To address these challenges, we propose ARCQuant, a framework that boosts NVFP4 performance via Augmented Residual Channels. Distinct from methods that compromise block isolation or hardware uniformity, ARCQuant maintains a strictly unified NVFP4 format by augmenting the activation matrix with quantized residual channels. This design integrates the error compensation process directly into the matrix reduction dimension, enabling the use of standard, highly optimized GEMM kernels with minimal overhead. Theoretical analysis confirms that the worst-case error bound of our dual-stage NVFP4 quantization is comparable to that of standard 8-bit formats such as MXFP8. Extensive experiments on LLaMA and Qwen models demonstrate that ARCQuant achieves state-of-the-art accuracy, comparable to full-precision baselines in perplexity and downstream tasks. Furthermore, deployment on RTX 5090 and RTX PRO 6000 GPUs confirms practical benefits, achieving up to 3x speedup over FP16. Our code is available at https://github.com/actypedef/ARCQuant. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Camouflaged Object Detection on CAMO (test) (S\_alpha, F\_w\_beta, E\_m, F\_m\_beta, MAE)](): Camouflaged Object Detection; CAMO (test); S\_alpha: 53.4. - [Camouflaged Object Detection on NC4K (S\_alpha, F\_m\_beta, MAE)](): Camouflaged Object Detection; NC4K; S\_alpha: 68.3. - [Camouflaged Object Detection on CHAMELEON](): Camouflaged Object Detection; Chameleon; S-measure (S\_alpha): 59.3. - [Camouflaged Object Detection on COD10K (test)](): Camouflaged Object Detection; COD10K (test); S-measure (S\_alpha): 0.64. ## Original sources and implementation [Original paper]() [PDF]()