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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Düzkar, Kemal
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918460904177664
author Düzkar, Kemal
author_facet Düzkar, Kemal
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2604_19758
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ThermoQA: A Three-Tier Benchmark for Evaluating Thermodynamic Reasoning in Large Language Models
Düzkar, Kemal
Artificial Intelligence
Computation and Language
Machine Learning
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
title ThermoQA: A Three-Tier Benchmark for Evaluating Thermodynamic Reasoning in Large Language Models
topic Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2604.19758