ReQAT: Achieving Full-Precision Reasoning Accuracy with 4-bit Floating-Point Quantization-Aware Training
On the map: LLM reasoning → Test-time compute and efficient thinking
Abstract (arXiv)
Large Reasoning Models (LRMs) achieve strong problem-solving through long chain-of-thought, but their deployment is constrained by the high cost of full-precision inference and growing KV cache footprints. Microscaled FP4 formats enable efficient FP4 deployment; however, fully quantizing weights, activations, and KV caches (W4A4KV4) causes severe reasoning degradation that existing PTQ and QAT fail to recover. We identify that FP4 failures concentrate on low-entropy tokens--precise symbolic commitments such as digits and operators--where quantization noise inflates sampling errors that cascade through reasoning traces. Based on this insight, we propose ReQAT, a reasoning-centric FP4 training framework with three components: (i) Trace-Aligned QAT (TAQ), which revisits identical reasoning traces to focus updates on critical low-entropy decisions; (ii) Selective Entropy Minimization (SEM), which reinforces confidence at low-entropy positions; and (iii) Q-FIT, a quantization-friendly initialization that jointly calibrates RoPE-consistent KV cache transformations to stabilize QAT. Under the same training budget, ReQAT not only recovers but surpasses BF16 fine-tuning accuracy, while delivering up to 3.9x throughput speedup on NVIDIA DGX Spark and 3.1x on B200.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 14 of its 50 references that are in the map. First posted on arXiv 14 Jun 2026.
Builds on
14 of 50 references.
Closest ancestors · 6
Same direction, newest first
- 2026A Case Study of Selected PTQ Baselines for Reasoning LLMs on Ascend NPU6 Feb
- 2026What Makes Low-Bit Quantization-Aware Training Work for Reasoning LLMs? A Systematic Study21 Jan
- 2025MixKVQ: Query-Aware Mixed-Precision KV Cache Quantization for Long-Context Reasoning22 Dec
- 2025gpt-oss-120b & gpt-oss-20b Model Card8 Aug
- 2025When Reasoning Meets Compression: Understanding the Effects of LLMs Compression on Large Reasoning Models2 Apr
- 2025Quantization Meets Reasoning: Exploring LLM Low-Bit Quantization Degradation for Mathematical Reasoning6 Jan
Other references · 8
- 2021Training Verifiers to Solve Math Word Problemscited by 10,109
- 2025Qwen3 Technical Reportcited by 8,820
- 2024OpenAI o1 System Cardcited by 1,244
- 2025OpenThoughts: Data Recipes for Reasoning Modelscited by 304
- 2023Microscaling Data Formats for Deep Learningcited by 152
- 2025Quartet: Native FP4 Training Can Be Optimal for Large Language Modelscited by 39
- 2025Training LLMs with MXFP4cited by 37
- 2025FireQ: Fast INT4-FP8 Kernel and RoPE-aware Quantization for LLM Inference Accelerationcited by 2
Built on by
Citing papers per year: 2026: 3.
3 papers.
Similar, not cited
Close in wording, no citation either way
- 2026ReSET: Accurate Latency-Critical NVFP4 Reasoning via Step-Aware Temperature Scaling
- 2026LAQuant: A Simple Overhead-free Large Reasoning Model Quantization by Layer-wise Lookahead Loss
- 2026BitCal-TTS: Bit-Calibrated Test-Time Scaling for Quantized Reasoning Models
- 2026QuantLRM: Quantization of Large Reasoning Models via Fine-Tuning Signals
- 2026RATIO: Reasoning Analysis and Token-level Inference Optimization for Quantized Reasoning Models
- 2025Quantization Hurts Reasoning? An Empirical Study on Quantized Reasoning Models
- 2026Attend to Your Own Thoughts: Breaking the Barrier for Post-Training Quantization of Reasoning LLMs through the Lens of 1.58-Bit Quantization
- 2026Extreme Low-Bit Inference in Reasoning Models: Failure Modes and Targeted Recovery