arXiv:2606.24391v1 Announce Type: cross 
Abstract: We introduce Age of LLM, a turn-based 1v1 benchmark in which two LLMs face off on a 13x7 grid to destroy the enemy base. Three stressors are deliberate: fog of war, full diplomacy (messages, ceasefires, ultimatums; uranium kept secret), and a reliability dimension where every turn must follow a strict JSON schema and an illegal action is silently discarded. The engine is private and each match uses a fresh random map seed and opponent, mitigating the data contamination that affects public benchmarks. Models receive a (near) rule-only prompt with no build-order advice (two tactical seed phrases were present during data collection; see Section 2.7). We benchmark 15 reasoning models across 54 matches and 5,258 actions. Findings: (1) the nuclear rush dominates (78% on the rules-coherent v0.11+ sub-corpus; 85% corpus-wide) with a sole-launcher signature that is largely mechanical under secret-simultaneous launch rules, not a cognitive deterrence failure; (2) military conquest is rare but faster (12.3 vs 18.9 turns); (3) diplomacy is prolific yet almost never consummated; (4) ~58% of illegal actions are fog/state errors, making the illegal-action rate a measure of belief-tracking; (5) -- the least established, and the only one we label exploratory -- a weak link associates reliability with winning. The corpus is small, unbalanced and not side-swapped, so the ranking is a preliminary descriptive view, not a contribution. Beyond ranking, the turn-by-turn traces of actions and messages make the corpus a lens on how LLMs reason under adversarial uncertainty -- their belief-tracking, spontaneous deception, and per-model cognitive "personas" -- which we frame as a future research direction. We release the replay format, an isometric viewer and all replays; engine source on request.

يقدم Age of LLM معيارًا استراتيجيًا 1v1 حيث يتنافس نموذجين للغة الكبيرة (LLMs) على شبكة 13x7 لتدمير قاعدة العدو، مع دمج عناصر مثل ضباب الحرب والدبلوماسية والالتزام الصارم بمخطط JSON للإجراءات. يهدف هذا المعيار إلى تقييم التفكير والموثوقية واتخاذ القرارات التكتيكية في LLMs تحت ظروف تنافسية.

Age of LLM presenta un benchmark estratégico 1v1 donde dos modelos de lenguaje de gran tamaño (LLMs) compiten en una cuadrícula de 13x7 para destruir una base enemiga, incorporando elementos como la niebla de guerra, la diplomacia y la estricta adherencia a un esquema JSON para las acciones. Este benchmark tiene como objetivo evaluar el razonamiento, la fiabilidad y la toma de decisiones tácticas en LLMs bajo condiciones competitivas.

L'Age of LLM introduit un benchmark stratégique 1v1 où deux modèles de langage de grande taille (LLMs) s'affrontent sur une grille de 13x7 pour détruire une base ennemie, intégrant des éléments tels que le brouillard de guerre, la diplomatie et le respect strict d'un schéma JSON pour les actions. Ce benchmark vise à évaluer le raisonnement, la fiabilité et la prise de décision tactique des LLMs dans des conditions compétitives.

The Age of LLM introduces a strategic 1v1 benchmark where two large language models (LLMs) compete on a 13x7 grid to destroy an enemy base, incorporating elements such as fog of war, diplomacy, and strict adherence to a JSON schema for actions. This benchmark aims to evaluate reasoning, reliability, and tactical decision-making in LLMs under competitive conditions.