PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning

TL;DR

PaCoRe introduces a parallel reasoning framework that significantly enhances test-time compute scalability in language models, enabling multi-million-token reasoning and surpassing existing systems in mathematics benchmarks.

Contribution

The paper presents a novel parallel coordinated reasoning framework that scales test-time compute beyond traditional sequential methods using message-passing and reinforcement learning.

Findings

Achieves 94.5% on HMMT 2025, surpassing GPT-5.

Scales effective TTC to roughly two million tokens.

Demonstrates strong improvements across diverse domains.

Abstract

We introduce Parallel Coordinated Reasoning (PaCoRe), a training-and-inference framework designed to overcome a central limitation of contemporary language models: their inability to scale test-time compute (TTC) far beyond sequential reasoning under a fixed context window. PaCoRe departs from the traditional sequential paradigm by driving TTC through massive parallel exploration coordinated via a message-passing architecture in multiple rounds. Each round launches many parallel reasoning trajectories, compacts their findings into context-bounded messages, and synthesizes these messages to guide the next round and ultimately produce the final answer. Trained end-to-end with large-scale, outcome-based reinforcement learning, the model masters the synthesis abilities required by PaCoRe and scales to multi-million-token effective TTC without exceeding context limits. The approach yields strong improvements across diverse domains, and notably pushes reasoning beyond frontier systems in mathematics: an 8B model reaches 94.5% on HMMT 2025, surpassing GPT-5's 93.2% by scaling effective TTC to roughly two million tokens. We open-source model checkpoints, training data, and the full inference pipeline to accelerate follow-up work.