---
format: "aidr-story-markdown/v1"
id: "2c44ff8988c41f3e362533336aa58a3180dc767b4b0537b587949829a8ce8a0b"
canonical_url: "https://aidr.today/2c44ff89?lang=en"
title: "LOLPlay: Benchmarking Real-Time Visual Agents in League of Legends"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-04T05:00:00.000Z"
category: "Research"
topics: ["openai","agent"]
source_urls: ["https://arxiv.org/abs/2611.03685"]
summary: "LOLPlay, a new benchmark for real-time visual agents in native League of Legends, evaluates 36 champion-specific tasks across resource management, combat, cooperation, and objectives using screenshots and mouse-keyboard actions. Across 324 task evaluations of three models at three reasoning-effort settings, GPT-6 Astra achieved the highest task success rate of 25.0% at low and medium effort and the highest subgoal success rate of 40.3% at medium effort, while high effort did not improve metrics over low effort for any model."
---

# LOLPlay: Benchmarking Real\-Time Visual Agents in League of Legends

> [Open the canonical story](<https://aidr.today/2c44ff89?lang=en>)

**Published:** 2026-11-04T05:00:00.000Z
**Category:** Research
**Topics:** openai, agent

## Summary

LOLPlay, a new benchmark for real\-time visual agents in native League of Legends, evaluates 36 champion\-specific tasks across resource management, combat, cooperation, and objectives using screenshots and mouse\-keyboard actions\. Across 324 task evaluations of three models at three reasoning\-effort settings, GPT\-6 Astra achieved the highest task success rate of 25\.0% at low and medium effort and the highest subgoal success rate of 40\.3% at medium effort, while high effort did not improve metrics over low effort for any model\.

## Sources

- [Story source](<https://arxiv.org/abs/2611.03685>)