# Explicit Reasoning Makes Better Judges: A Systematic Study on Accuracy, Efficiency, and Robustness

arXiv: https://arxiv.org/abs/2509.13332 · HTML: https://mlsanity.com/paper/2509.13332/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 23 of its 62 references that are in the map.
- First posted on arXiv 9 Sep 2025.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)
- LLM alignment / RLHF / preference optimisation / [Process reward models for reasoning](https://mlsanity.com/alignment/process-reward-models-for-reasoning/)

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

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models](https://mlsanity.com/paper/2503.09567/) (2025)
- [Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge](https://mlsanity.com/paper/2410.02736/) (2024)
- [Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Models](https://mlsanity.com/paper/2501.09686/) (2025)
- [Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling](https://mlsanity.com/paper/2502.06703/) (2025)
- [Optimizing Test-Time Compute via Meta Reinforcement Fine-Tuning](https://mlsanity.com/paper/2503.07572/) (2025)
- [Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning](https://mlsanity.com/paper/2502.18080/) (2025)
- [A Survey of Frontiers in LLM Reasoning: Inference Scaling, Learning to Reason, and Agentic Systems](https://mlsanity.com/paper/2504.09037/) (2025)
- [Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoning](https://mlsanity.com/paper/2412.09078/) (2024)
- [Adaptive Inference-Time Compute: LLMs Can Predict if They Can Do Better, Even Mid-Generation](https://mlsanity.com/paper/2410.02725/) (2024)
- [Efficient Inference for Large Reasoning Models: A Survey](https://mlsanity.com/paper/2503.23077/) (2025)
- [A Survey of Test-Time Compute: From Intuitive Inference to Deliberate Reasoning](https://mlsanity.com/paper/2501.02497/) (2025)
- [CountQA: How Well Do MLLMs Count in the Wild?](https://mlsanity.com/paper/2508.06585/) (2025)
- [GeoChain: Multimodal Chain-of-Thought for Geographic Reasoning](https://mlsanity.com/paper/2506.00785/) (2025)
- [Bag of Tricks for Inference-time Computation of LLM Reasoning](https://mlsanity.com/paper/2502.07191/) (2025)
- [The Energy Cost of Reasoning: Analyzing Energy Usage in LLMs with Test-time Compute](https://mlsanity.com/paper/2505.14733/) (2025)
- [MAEA: Multimodal Attribution for Embodied AI](https://mlsanity.com/paper/2307.13850/) (2023)
- [HueManity: Probing Fine-Grained Visual Perception in MLLMs](https://mlsanity.com/paper/2506.03194/) (2025)
- [Semantic Augmentation in Images using Language](https://mlsanity.com/paper/2404.02353/) (2024)
- [Attribution Regularization for Multimodal Paradigms](https://mlsanity.com/paper/2404.02359/) (2024)
- [DynamicMind: A Tri-Mode Thinking System for Large Language Models](https://mlsanity.com/paper/2506.05936/) (2025)
- [Retraining DistilBERT for a Voice Shopping Assistant by Using Universal Dependencies](https://mlsanity.com/paper/2103.15737/) (2021)

## Built on by (most cited first, 3 of 3)

- [From Holistic Evaluation to Structured Criteria: Rubrics Across the Evolving LLM Landscape](https://mlsanity.com/paper/2606.08625/) (2026)
- [Do We Need Frontier Models to Verify Mathematical Proofs?](https://mlsanity.com/paper/2604.02450/) (2026)
- [SLMJury: Can Small Language Models Judge as Well as Large Ones?](https://mlsanity.com/paper/2606.07810/) (2026)

## Abstract (arXiv)

As Large Language Models (LLMs) are increasingly adopted as automated judges in benchmarking and reward modeling, ensuring their reliability, efficiency, and robustness has become critical. In this work, we present a systematic comparison of "thinking" and "non-thinking" LLMs in the LLM-as-a-judge paradigm using open-source Qwen 3 models of relatively small sizes (0.6B, 1.7B, and 4B parameters). We evaluate both accuracy and computational efficiency (FLOPs) on RewardBench tasks, and further examine augmentation strategies for non-thinking models, including in-context learning, rubric-guided judging, reference-based evaluation, and n-best aggregation. Our results show that despite these enhancements, non-thinking models generally fall short of their thinking counterparts. Our results show that thinking models achieve approximately 10% points higher accuracy with little overhead (under 2x), in contrast to augmentation strategies like few-shot learning, which deliver modest gains at a higher cost (>8x). Bias and robustness analyses further demonstrate that thinking models maintain significantly greater consistency under a variety of bias conditions such as positional, bandwagon, identity, diversity, and random biases (6% higher on average). We further extend our experiments to the multilingual setting and our results confirm that explicit reasoning extends its benefits beyond English. Overall, our work results in several important findings that provide systematic evidence that explicit reasoning offers clear advantages in the LLM-as-a-judge paradigm not only in accuracy and efficiency but also in robustness.