# Tequila: Trapping-free Ternary Quantization for Large Language Models

arXiv: https://arxiv.org/abs/2509.23809 · HTML: https://mlsanity.com/paper/2509.23809/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 36 references that are in the map.
- First posted on arXiv 28 Sep 2025.

## Builds on (most cited first, 27 of 27)

- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [PIQA: Reasoning about Physical Commonsense in Natural Language](https://mlsanity.com/paper/1911.11641/) (2019)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models](https://mlsanity.com/paper/2211.10438/) (2022)
- [LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale](https://mlsanity.com/paper/2208.07339/) (2022)
- [Learned Step Size Quantization](https://mlsanity.com/paper/1902.08153/) (2019)
- [8-bit Optimizers via Block-wise Quantization](https://mlsanity.com/paper/2110.02861/) (2021)
- [A Survey on Large Language Model (LLM) Security and Privacy: The Good, the Bad, and the Ugly](https://mlsanity.com/paper/2312.02003/) (2023)
- [Trained Ternary Quantization](https://mlsanity.com/paper/1612.01064/) (2016)
- [Ternary Weight Networks](https://mlsanity.com/paper/1605.04711/) (2016)
- [BitNet: Scaling 1-bit Transformers for Large Language Models](https://mlsanity.com/paper/2310.11453/) (2023)
- [QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models](https://mlsanity.com/paper/2310.08041/) (2023)
- [Extremely Low Bit Neural Network: Squeeze the Last Bit Out with ADMM](https://mlsanity.com/paper/1707.09870/) (2017)
- [MiniCPM4: Ultra-Efficient LLMs on End Devices](https://mlsanity.com/paper/2506.07900/) (2025)
- [ParetoQ: Improving Scaling Laws in Extremely Low-bit LLM Quantization](https://mlsanity.com/paper/2502.02631/) (2025)
- [BitNet b1.58 2B4T Technical Report](https://mlsanity.com/paper/2504.12285/) (2025)
- [Ultra-FineWeb: Efficient Data Filtering and Verification for High-Quality LLM Training Data](https://mlsanity.com/paper/2505.05427/) (2025)
- [AlphaTuning: Quantization-Aware Parameter-Efficient Adaptation of Large-Scale Pre-Trained Language Models](https://mlsanity.com/paper/2210.03858/) (2022)
- [Bitnet.cpp: Efficient Edge Inference for Ternary LLMs](https://mlsanity.com/paper/2502.11880/) (2025)
- [TernaryLLM: Ternarized Large Language Model](https://mlsanity.com/paper/2406.07177/) (2024)
- [Quaff: Quantized Parameter-Efficient Fine-Tuning under Outlier Spatial Stability Hypothesis](https://mlsanity.com/paper/2505.14742/) (2025)
- [Ternary Quantization: A Survey](https://mlsanity.com/paper/2303.01505/) (2023)

## Built on by (most cited first, 9 of 9)

- [DuQuant++: Fine-grained Rotation Enhances Microscaling FP4 Quantization](https://mlsanity.com/paper/2604.17789/) (2026)
- [AngelSlim: A more accessible, comprehensive, and efficient toolkit for large model compression](https://mlsanity.com/paper/2602.21233/) (2026)
- [GSQ: Highly-Accurate Low-Precision Scalar Quantization for LLMs via Gumbel-Softmax Sampling](https://mlsanity.com/paper/2604.18556/) (2026)
- [CAT-Q: Cost-efficient and Accurate Ternary Quantization for LLMs](https://mlsanity.com/paper/2606.26650/) (2026)
- [Attend to Your Own Thoughts: Breaking the Barrier for Post-Training Quantization of Reasoning LLMs through the Lens of 1.58-Bit Quantization](https://mlsanity.com/paper/2608.01078/) (2026)
- [Breaking the 1.58-bit Barrier for Ternary LLMs](https://mlsanity.com/paper/2609.16338/) (2026)
- [PPMStereo: Pick-and-Play Memory Construction for Consistent Dynamic Stereo Matching](https://mlsanity.com/paper/2510.20178/) (2025)
- [HESTIA: A Hessian-Guided Differentiable Quantization-Aware Training Framework for Extremely Low-Bit LLMs](https://mlsanity.com/paper/2601.20745/) (2026)
- [Lost in the bf16 Cast: Exporting Ternary Language Models Can Revert Most Low-Learning-Rate Code Changes](https://mlsanity.com/paper/2610.07853/) (2026)

## Abstract (arXiv)

Quantization techniques are essential for the deployment of Large Language Models (LLMs) on edge devices. However, prevailing methods often rely on mixed-precision multiplication that lacks efficient hardware support, making it not feasible. Ternary weight quantization addresses this by constraining weights to {-1, 0, 1}, replacing expensive multiplications with hardware-efficient additions. However, such aggressive compression leads to significant accuracy degradation, even after costly quantization-aware training with massive data. We identify the core issue as deadzone trapping: a large number of weights are trapped at the deadzone boundary. This occurs because these weights receive only noisy, uninformative gradients, preventing stable escape from the deadzone and severely impeding model capacity and optimization. To address this issue, we propose Tequila, a trapping-free quantization optimization method that reactivates deadzone-trapped weights by repurposing them as dynamic biases. This allows the repurposed weights to provide a continuous signal in the forward pass and, critically, receive direct, meaningful gradient signals during backpropagation, thereby enhancing model capacity and optimization with nearly zero inference overhead. Extensive evaluations demonstrate that Tequila outperforms state-of-the-art (SOTA) ternary quantization methods across five benchmarks. Specifically, on the ARC benchmark, it achieves >4% accuracy gain over the SOTA baseline, nearly matching full-precision performance (within <1% gap) with a 3.0x inference speedup. Consequently, Tequila offers a highly practical and efficient implementation for the deployment of advanced LLMs in resource-constrained environments. The code is available at https://github.com/Tencent/AngelSlim.