# The Silent Hyperparameter: Quantifying the Impact of Inference Backends on LLM Reproducibility

arXiv: https://arxiv.org/abs/2605.19537 · HTML: https://mlsanity.com/paper/2605.19537/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 30 references that are in the map.
- First posted on arXiv 19 May 2026.

## Builds on (most cited first, 23 of 23)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code](https://mlsanity.com/paper/2403.07974/) (2024)
- [SGLang: Efficient Execution of Structured Language Model Programs](https://mlsanity.com/paper/2312.07104/) (2023)
- [JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models](https://mlsanity.com/paper/2404.01318/) (2024)
- [Lessons from the Trenches on Reproducible Evaluation of Language Models](https://mlsanity.com/paper/2405.14782/) (2024)
- [A Thorough Examination of Decoding Methods in the Era of LLMs](https://mlsanity.com/paper/2402.06925/) (2024)
- [Towards Efficient Generative Large Language Model Serving: A Survey from Algorithms to Systems](https://mlsanity.com/paper/2312.15234/) (2023)
- [SimpleQA Verified: A Reliable Factuality Benchmark to Measure Parametric Knowledge](https://mlsanity.com/paper/2509.07968/) (2025)
- [Understanding and Mitigating Numerical Sources of Nondeterminism in LLM Inference](https://mlsanity.com/paper/2506.09501/) (2025)
- [Large Language Model Inference Acceleration: A Comprehensive Hardware Perspective](https://mlsanity.com/paper/2410.04466/) (2024)
- [Benchmarking Prompt Sensitivity in Large Language Models](https://mlsanity.com/paper/2502.06065/) (2025)
- ["Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization](https://mlsanity.com/paper/2411.02355/) (2024)
- [LLM-Inference-Bench: Inference Benchmarking of Large Language Models on AI Accelerators](https://mlsanity.com/paper/2411.00136/) (2024)
- [Towards Reproducible LLM Evaluation: Quantifying Uncertainty in LLM Benchmark Scores](https://mlsanity.com/paper/2410.03492/) (2024)
- [Chasing Shadows: Pitfalls in LLM Security Research](https://mlsanity.com/paper/2512.09549/) (2025)
- [A Survey on Inference Engines for Large Language Models: Perspectives on Optimization and Efficiency](https://mlsanity.com/paper/2505.01658/) (2025)
- [Hardware Acceleration for Neural Networks: A Comprehensive Survey](https://mlsanity.com/paper/2512.23914/) (2025)
- [mlf-core: a framework for deterministic machine learning](https://mlsanity.com/paper/2104.07651/) (2021)
- [Hidden Reliability Risks in Large Language Models: Systematic Identification of Precision-Induced Output Disagreements](https://mlsanity.com/paper/2604.19790/) (2026)

## Built on by (most cited first, 8 of 8)

- [The Integer Alibi: Localizing Cross-Kernel Divergence in INT8-Quantized LLM Inference](https://mlsanity.com/paper/2608.13756/) (2026)
- [IBIB: A Protocol for Measuring Enterprise AI Systems by Serving Route, Not Model Identifier](https://mlsanity.com/paper/2609.10494/) (2026)
- [Grounded Inference: Principles for Deterministically Encapsulated Generative Models](https://mlsanity.com/paper/2606.19753/) (2026)
- [Same Request, Different Answer: Quantization Amplifies Cache-Induced Divergence in LLM Serving](https://mlsanity.com/paper/2609.04748/) (2026)
- [Measuring the Serving Stack Instead of the Model: Hidden Confounds in Local Tool-Use Evaluation](https://mlsanity.com/paper/2609.26693/) (2026)
- [RECLAIM: Can Agents Reproduce the Claims of Machine Learning Papers?](https://mlsanity.com/paper/2609.28850/) (2026)
- [What Does a Harness Repair? A Preregistered Study of Visibility, Baseline Adequacy and Evaluation Defects](https://mlsanity.com/paper/2610.05533/) (2026)
- [Same Text, Different Prediction: Serving-Context Nondeterminism in Text Classifiers](https://mlsanity.com/paper/2610.09111/) (2026)

## Abstract (arXiv)

Progress in LLMs is increasingly measured through standardized benchmarks, where state-of-the-art improvements are often separated by fractions of a percentage point. At the same time, the computational cost of evaluating modern LLMs has driven widespread adoption of specialized inference backends, software systems that execute trained models efficiently at inference time. While critical for scalability, system-level optimizations, such as custom CUDA kernels and reduced-precision arithmetic, can alter token probabilities and introduce non-determinism, possibly cascading into divergent generation. In this work, we first survey the inference landscape, identifying 200 distinct engines, and analyze 35,000 ML publications, finding that the specific inference stack is rarely reported despite this widespread diversity. We then present a systematic empirical study of how inference backends affect LLM benchmark results. Holding model weights, decoding parameters, and hardware constant, we evaluate five widely used inference engines, including vLLM, SGLang, and llama.cpp, across multiple open-weight models and established benchmarks. We show that the choice of backend alone can shift benchmark scores by up to 16.6 percentage points and induce high rates of output disagreement. By isolating backend optimizations and tracing the execution pipeline, we find this divergence is driven by system-level optimizations like prefix caching and CUDA graphs, custom kernels, and engine-specific defaults in logit processing. Our findings identify the inference backend as a previously unreported but consequential hyperparameter in the evaluation of LLM and advocate standardized reporting of inference stacks to improve the reproducibility and interpretability of benchmark comparisons.