# ThermoQA: A Three-Tier Benchmark for Evaluating Thermodynamic Reasoning in Large Language Models

arXiv: https://arxiv.org/abs/2604.19758 · HTML: https://mlsanity.com/paper/2604.19758/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 8 of its 18 references that are in the map.
- First posted on arXiv 25 Mar 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 8 of 8)

- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [SciBench: Evaluating College-Level Scientific Problem-Solving Abilities of Large Language Models](https://mlsanity.com/paper/2307.10635/) (2023)
- [MatSciBench: Benchmarking the Reasoning Ability of Large Language Models in Materials Science](https://mlsanity.com/paper/2510.12171/) (2025)
- [EngiBench: A Benchmark for Evaluating Large Language Models on Engineering Problem Solving](https://mlsanity.com/paper/2509.17677/) (2025)
- [Using Large Language Models for Solving Thermodynamic Problems](https://mlsanity.com/paper/2502.05195/) (2025)
- [From Canonical to Complex: Benchmarking LLM Capabilities in Undergraduate Thermodynamics](https://mlsanity.com/paper/2508.21452/) (2025)
- [Superstudent intelligence in thermodynamics](https://mlsanity.com/paper/2506.09822/) (2025)

## Abstract (arXiv)

We present ThermoQA, a benchmark of 293 open-ended engineering thermodynamics problems in three tiers: property lookups (110 Q), component analysis (101 Q), and full cycle analysis (82 Q). Ground truth is computed programmatically from CoolProp 7.2.0, covering water, R-134a, and variable-cp air. Six frontier LLMs are evaluated across three independent runs each. The composite leaderboard is led by Claude Opus 4.6 (94.1%), GPT-5.4 (93.1%), and Gemini 3.1 Pro (92.5%). Cross-tier degradation ranges from 2.8 pp (Opus) to 32.5 pp (MiniMax), confirming that property memorization does not imply thermodynamic reasoning. Supercritical water, R-134a refrigerant, and combined-cycle gas turbine analysis serve as natural discriminators with 40-60 pp performance spreads. Multi-run sigma ranges from +/-0.1% to +/-2.5%, quantifying reasoning consistency as a distinct evaluation axis. Dataset and code are open-source at https://huggingface.co/datasets/olivenet/thermoqa