# NeutronGym: Physics-Graded Neutron Instrument Design for LLM Agents

arXiv: https://arxiv.org/abs/2610.03631 · HTML: https://mlsanity.com/paper/2610.03631/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Educational agents and math problem solving, one of 16 directions in LLM agents and tool use.
- Builds on 43 of its 84 references that are in the map.
- First posted on arXiv 2 Oct 2026.

## Placed in

- LLM agents and tool use / [Educational agents and math problem solving](https://mlsanity.com/agents/educational-agents-and-math-problem-solving/)

## Builds on (most cited first, 30 of 43)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [STaR: Bootstrapping Reasoning With Reasoning](https://mlsanity.com/paper/2203.14465/) (2022)
- [RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment](https://mlsanity.com/paper/2304.06767/) (2023)
- [YaRN: Efficient Context Window Extension of Large Language Models](https://mlsanity.com/paper/2309.00071/) (2023)
- [Beyond Human Data: Scaling Self-Training for Problem-Solving with Language Models](https://mlsanity.com/paper/2312.06585/) (2023)
- [Training Software Engineering Agents and Verifiers with SWE-Gym](https://mlsanity.com/paper/2412.21139/) (2024)
- [ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery](https://mlsanity.com/paper/2410.05080/) (2024)
- [ReST meets ReAct: Self-Improvement for Multi-Step Reasoning LLM Agent](https://mlsanity.com/paper/2312.10003/) (2023)
- [Feedback Loops With Language Models Drive In-Context Reward Hacking](https://mlsanity.com/paper/2402.06627/) (2024)
- [RLVE: Scaling Up Reinforcement Learning for Language Models with Adaptive Verifiable Environments](https://mlsanity.com/paper/2511.07317/) (2025)
- [El Agente: An Autonomous Agent for Quantum Chemistry](https://mlsanity.com/paper/2505.02484/) (2025)
- [GenEnv: Difficulty-Aligned Co-Evolution Between LLM Agents and Environment Simulators](https://mlsanity.com/paper/2512.19682/) (2025)
- [Reward Hacking Benchmark: Measuring Exploits in LLM Agents with Tool Use](https://mlsanity.com/paper/2605.02964/) (2026)
- [LLMs Gaming Verifiers: RLVR can Lead to Reward Hacking](https://mlsanity.com/paper/2604.15149/) (2026)
- [GEM: A Gym for Agentic LLMs](https://mlsanity.com/paper/2510.01051/) (2025)
- [Toward Engineering AGI: Benchmarking the Engineering Design Capabilities of LLMs](https://mlsanity.com/paper/2509.16204/) (2025)
- [BoxingGym: Benchmarking Progress in Automated Experimental Design and Model Discovery](https://mlsanity.com/paper/2501.01540/) (2025)
- [MooseAgent: A LLM Based Multi-agent Framework for Automating Moose Simulation](https://mlsanity.com/paper/2504.08621/) (2025)
- [Frontier-Eng: Benchmarking Self-Evolving Agents on Real-World Engineering Tasks with Generative Optimization](https://mlsanity.com/paper/2604.12290/) (2026)
- [SciAgentGym: Benchmarking Multi-Step Scientific Tool-use in LLM Agents](https://mlsanity.com/paper/2602.12984/) (2026)
- [Hack-Verifiable Environments: Towards Evaluating Reward Hacking at Scale](https://mlsanity.com/paper/2605.20744/) (2026)
- [BeyondBench: Contamination-Resistant Evaluation of Reasoning in Language Models](https://mlsanity.com/paper/2509.24210/) (2025)
- [MDAgent2: Large Language Model for Code Generation and Knowledge Q&A in Molecular Dynamics](https://mlsanity.com/paper/2601.02075/) (2026)
- [Collider-Bench: Benchmarking AI Agents with Particle Physics Analysis Reproduction](https://mlsanity.com/paper/2605.13950/) (2026)
- [Before the Model Learns the Bug:Fuzzing RLVR Verifiers](https://mlsanity.com/paper/2606.01066/) (2026)
- [GRACE: an Agentic AI for Particle Physics Experiment Design and Simulation](https://mlsanity.com/paper/2602.15039/) (2026)
- [When Is an LLM Worth It for Hyperparameter Optimization? A Budget-Matched Study on Tabular Data Finds the Warm-Start Is a Default Configuration, Not the Model](https://mlsanity.com/paper/2606.21641/) (2026)
- [MDGYM: Benchmarking AI Agents on Molecular Simulations](https://mlsanity.com/paper/2605.08941/) (2026)

## Abstract (arXiv)

Designing a scientific instrument tests whether language-model agents can do physics rather than recall it, provided the grading cannot be argued with. We introduce NeutronGym, to our knowledge the first executable environment for neutron instrument design: agents build instruments through validating tools, McStas ray-traces what they build, and a level-resolved ladder grades syntax, runtime, structure and science with no LLM judge. Procedural families supply unlimited instances of a fixed layout whose design parameters the agent must set, with held-out parameter regimes; a curated slice, McStasBench, adds 16 tasks from published instruments behind memorization probes and a sandbox. Seven models reproduce at most 7 of the 16, none retrieves a reference, and none meets an improvement target. The environment also trains. Reinforcement learning on its reward takes Qwen3-8B from 11% to 77% of held-out instances of a family whose targets come from a hidden design (69% at a second seed), past an untrained Qwen3-32B, and the recipe holds, at one seed each, on three further gated families. The analysis says what that gain is. Without the ladder's partial credit it collapses by 60 points. From reward alone the trained model reaches what a classical optimizer reaches, at the agent's simulation budget, only when handed the closed-form physics (77% against 81%, a gap that does not separate at this size), while frontier models still solve 98-99%. Getting a trustworthy result meant failing four task designs that no-model baselines could solve, and we release the probes that found them.