EdgeRunner 20B: Military Task Parity with GPT-5 while Running on the Edge
Abstract
We present EdgeRunner 20B, a fine-tuned version of gpt-oss-20b optimized for military tasks. EdgeRunner 20B was trained on 1.6M high-quality records curated from military documentation and websites. We also present four new tests sets: (a) combat arms, (b) combat medic, (c) cyber operations, and (d) mil-bench-5k (general military knowledge). On these military test sets, EdgeRunner 20B matches or exceeds GPT-5 task performance with 95%+ statistical significance, except for the high reasoning setting on the combat medic test set and the low reasoning setting on the mil-bench-5k test set. Versus gpt-oss-20b, there is no statistically-significant regression on general-purpose benchmarks like ARC-C, GPQA Diamond, GSM8k, IFEval, MMLU Pro, or TruthfulQA, except for GSM8k in the low reasoning setting. We also present analyses on hyperparameter settings, cost, and throughput. These findings show that small, locally-hosted models are ideal solutions for data-sensitive operations such as in the military domain, allowing for deployment in air-gapped edge devices.
Cite
@article{arxiv.2510.26550,
title = {EdgeRunner 20B: Military Task Parity with GPT-5 while Running on the Edge},
author = {Jack FitzGerald and Aristotelis Lazaridis and Dylan Bates and Aman Sharma and Jonnathan Castillo and Yousif Azami and Sean Bailey and Jeremy Cao and Peter Damianov and Kevin de Haan and Luke Kerbs and Vincent Lu and Joseph Madigan and Jeremy McLaurin and Jonathan Tainer and Dave Anderson and Jonathan Beck and Jamie Cuticello and Colton Malkerson and Tyler Saltsman},
journal= {arXiv preprint arXiv:2510.26550},
year = {2025}
}
Comments
19 pages; v2 includes an additional appendix with test set examples