Reward Reasoning Models

Reward models play a critical role in guiding large language models toward outputs that align with human expectations. However, an open challenge remains in effectively utilizing test-time compute to enhance reward model performance. In this work, we introduce Reward Reasoning Models (RRMs), which are specifically designed to execute a deliberate reasoning process before generating final rewards. Through chain-of-thought reasoning, RRMs leverage additional test-time compute for complex queries where appropriate rewards are not immediately apparent. To develop RRMs, we implement a reinforcement learning framework that fosters selfevolved reward reasoning capabilities without requiring explicit reasoning traces as training data. Experimental results demonstrate that RRMs achieve superior performance on reward modeling benchmarks across diverse domains. Notably, we show that RRMs can adaptively exploit test-time compute to further improve reward accuracy. The pretrained models are available at https://huggingface. co/Reward-Reasoning.

Frontier Reward Modeling Performance Figure 1: Average accuracy of various reward models on Preference Proxy Evaluations [18] over the MMLU-Pro, MATH, and GPQA subsets. The proposed reward reasoning model (RRM) outperforms previous reward models across model sizes. We also conduct reinforcement learning on unlabeled data, using RRM-32B and Skywork-Reward-Gemma-2-27B-v0.2 [40] as the reward models, respectively. Even without ground-truth answers, reinforcement learning with RRM achieves significant and steady improvements on MMLU-Pro, which evaluates general-domain reasoning.