AndroidWorld

vision

AndroidWorld evaluates an agent's ability to operate in real Android GUI environments, completing multi-step tasks by perceiving screen content and executing touch/type actions.

Methodology

Imported from llm-stats public benchmark metadata. Modality: multimodal. Max score: 1. Categories: agents, vision. Language: en. Verified by llm-stats: no.

Leaderboard

  1. GLM-5V-Turbo self-reported llm-stats
    75.7%
  2. Qwen3.6-27B self-reported llm-stats
    70.3%