# LLM-Cave: A benchmark and light environment for large language models reasoning and decision-making system

arXiv: https://arxiv.org/abs/2511.22598 · HTML: https://mlsanity.com/paper/2511.22598/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 16 of its 20 references that are in the map.
- First posted on arXiv 27 Nov 2025.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 16 of 16)

- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Generative Agents: Interactive Simulacra of Human Behavior](https://mlsanity.com/paper/2304.03442/) (2023)
- [ChatDev: Communicative Agents for Software Development](https://mlsanity.com/paper/2307.07924/) (2023)
- [LM-Nav: Robotic Navigation with Large Pre-Trained Models of Language, Vision, and Action](https://mlsanity.com/paper/2207.04429/) (2022)
- [Chain of Draft: Thinking Faster by Writing Less](https://mlsanity.com/paper/2502.18600/) (2025)
- [TradingAgents: Multi-Agents LLM Financial Trading Framework](https://mlsanity.com/paper/2412.20138/) (2024)
- [Large Language Models Play StarCraft II: Benchmarks and A Chain of Summarization Approach](https://mlsanity.com/paper/2312.11865/) (2023)
- [Can an Embodied Agent Find Your "Cat-shaped Mug"? LLM-Guided Exploration for Zero-Shot Object Navigation](https://mlsanity.com/paper/2303.03480/) (2023)
- [Game-theoretic LLM: Agent Workflow for Negotiation Games](https://mlsanity.com/paper/2411.05990/) (2024)
- [RobotGPT: Robot Manipulation Learning from ChatGPT](https://mlsanity.com/paper/2312.01421/) (2023)
- [LLaMA Rider: Spurring Large Language Models to Explore the Open World](https://mlsanity.com/paper/2310.08922/) (2023)
- [SwarmBrain: Embodied agent for real-time strategy game StarCraft II via large language models](https://mlsanity.com/paper/2401.17749/) (2024)
- [Enhancing the LLM-Based Robot Manipulation Through Human-Robot Collaboration](https://mlsanity.com/paper/2406.14097/) (2024)
- [LLM-PySC2: Starcraft II learning environment for Large Language Models](https://mlsanity.com/paper/2411.05348/) (2024)
- [LLM-Enhanced Path Planning: Safe and Efficient Autonomous Navigation with Instructional Inputs](https://mlsanity.com/paper/2412.02655/) (2024)
- [Reflection of Episodes: Learning to Play Game from Expert and Self Experiences](https://mlsanity.com/paper/2502.13388/) (2025)

## Abstract (arXiv)

Large language models (LLMs) such as ChatGPT o1, ChatGPT o3, and DeepSeek R1 have shown great potential in solving difficult problems. However, current LLM evaluation benchmarks are limited to one-step interactions. Some of the existing sequence decision-making environments, such as TextStarCraftII and LLM-PySC2, are too complicated and require hours of interaction to complete a game. In this paper, we introduce LLM-Cave, a benchmark and light environment for LLM reasoning and decision-making systems. This environment is a classic instance in the era of Symbolism. Artificial intelligence enables the agent to explore the environment and avoid potential losses by reasoning about nearby dangers using partial observable state information. In the experiment, we evaluated the sequential reasoning ability, decision-making performance and computational efficiency of mainstream large language models (LLMs) such as GPT-4o-mini, o1-mini, and DeepSeek-R1. Experiments show that while Deepseek-R1 achieved the highest success rate on complex reasoning tasks, smaller models like 4o-mini significantly narrowed the performance gap on challenges by employing Chain of Speculation and Planner-Critic strategies, at the expense of reduced computational efficiency. This indicates that structured, multi-step reasoning combined with an LLM-based feedback mechanism can substantially enhance an LLM's decision-making capabilities, providing a promising direction for improving reasoning in weaker models and suggesting a new reasoning-centered benchmark for LLM assessment. Our code is open-sourced in https://github.com/puleya1277/CaveEnv.