# World Potential Model: Pretrained World Knowledge as Progress Potentials

arXiv: https://arxiv.org/abs/2610.09560 · HTML: https://mlsanity.com/paper/2610.09560/

## Key facts

- No citing papers in the map yet (data through 7 Oct 2026).
- Builds on 12 of its 42 references that are in the map.
- First posted on arXiv 7 Oct 2026.

## Builds on (most cited first, 12 of 12)

- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [ART: Automatic multi-step reasoning and tool-use for large language models](https://mlsanity.com/paper/2303.09014/) (2023)
- [Process Reward Models for LLM Agents: Practical Framework and Directions](https://mlsanity.com/paper/2502.10325/) (2025)
- [AgentPRM: Process Reward Models for LLM Agents via Step-Wise Promise and Progress](https://mlsanity.com/paper/2511.08325/) (2025)
- [Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering](https://mlsanity.com/paper/2604.08224/) (2026)
- [Hindsight Credit Assignment for Long-Horizon LLM Agents](https://mlsanity.com/paper/2603.08754/) (2026)
- [AgencyBench: Benchmarking the Frontiers of Autonomous Agents in 1M-Token Real-World Contexts](https://mlsanity.com/paper/2601.11044/) (2026)
- [ProgRM: Build Better GUI Agents with Progress Rewards](https://mlsanity.com/paper/2505.18121/) (2025)
- [A Subgoal-driven Framework for Improving Long-Horizon LLM Agents](https://mlsanity.com/paper/2603.19685/) (2026)
- [WisPaper: Your AI Scholar Search Engine](https://mlsanity.com/paper/2512.06879/) (2025)
- [LH-Bench: Skill-Grounded Evaluation of Long-Horizon Agents on Subjective Enterprise Tasks](https://mlsanity.com/paper/2603.22744/) (2026)
- [Automating Potential-based Reward Shaping with Vision Language Model Guidance](https://mlsanity.com/paper/2606.27180/) (2026)

## Abstract (arXiv)

Long-horizon language agents often receive supervision only from terminal task outcomes, leaving little signal for distinguishing productive intermediate behavior from stagnation or even regression. Rather than learning a separate value function or process reward model for every task, we ask whether pretrained models can recognize task progress from their existing world knowledge. We formalize this capability with a World Potential Model (WPM), a goal-conditioned evaluator of task-relative realized progress in agent contexts. In ALFWorld and ScienceWorld, off-the-shelf pretrained models substantially outperform chance at recovering realized-progress structure without task-specific evaluator fine-tuning. We further anchor these progress judgments to task-specific milestones to obtain scalar world potentials, whose temporal differences provide process-sensitive step-level credit for policy optimization. Under matched comparisons, WPM-guided optimization improves success over outcome-only GRPO across all evaluated configurations. Together, these results provide initial evidence that pretrained world knowledge can support reusable realized-progress evaluation and provide useful supervision for long-horizon agents.