# The Bidirectional Process Reward Model

Source: [The Bidirectional Process Reward Model](<https://www.sota2.com/research/paper/68fa7a681e209131cea83008>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-07-02T05:00:05.659Z

Authors: Lingyin Zhang, Jun Gao, Xiaoxue Ren, Ziqiang Cao

Published: 2025-08-03T09:23:49.000Z

## Abstract

Process Reward Models (PRMs), which assign fine-grained scores to intermediate reasoning steps within a solution trajectory, have emerged as a promising approach to enhance the reasoning quality of Large Language Models (LLMs). However, most existing PRMs rely on a unidirectional left-to-right (L2R) evaluation scheme, which restricts their utilization of global context. In light of this challenge, we propose a novel bidirectional evaluation paradigm, named Bidirectional Process Reward Model (BiPRM). BiPRM incorporates a parallel right-to-left (R2L) evaluation stream, implemented via prompt reversal, alongside the conventional L2R flow. Then a gating mechanism is introduced to adaptively fuse the reward scores from both streams to yield a holistic quality assessment. Remarkably, compared to the original PRM, BiPRM introduces only a 0.3% parameter increase for the gating module, and the parallel execution of two streams incurs merely 5% inference time latency. Our extensive empirical evaluations spanning diverse benchmarks, LLM backbones, PRM objectives and sampling policies demonstrate that BiPRM consistently surpasses unidirectional baselines, achieving an average relative gain of 10.6% over 54 solution-level configurations and 37.7% in 12 step-level error detection scenarios. Generally, our results highlight the effectiveness, robustness and general applicability of BiPRM, offering a promising new direction for process-based reward modeling.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Mathematical Reasoning on MATH500 (BCE, MSE, Q-value rankings)](<https://www.sota2.com/research/sota/mathematical-reasoning-on-math500-bce-mse-q-value-rankings.md>): Mathematical Reasoning; MATH500; BCE: 21.

- [Process Reward Modeling on ProcessBench](<https://www.sota2.com/research/sota/process-reward-modeling-on-processbench.md>): Process Reward Modeling; PROCESSBENCH; GSM8K Accuracy: 46.8.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2508.01682>)

[PDF](<https://arxiv.org/pdf/2508.01682.pdf>)