AndroidWorld_SR
AndroidWorld Success Rate (SR) benchmark - A dynamic benchmarking environment for autonomous agents operating on Android devices. Evaluates agents on 116 programmatic tasks across 20 real-world Android apps using multimodal inputs (screen screenshots, accessibility trees, and natural language instructions). Measures success rate of agents completing tasks like sending messages, creating calendar events, and navigating mobile interfaces. Published at ICLR 2025. Best current performance: 30.6% success rate (M3A agent) vs 80.0% human performance.
Qwen3.5-35B-A3B from Alibaba Cloud / Qwen Team currently leads the AndroidWorld_SR leaderboard with a score of 0.711 across 8 evaluated AI models.
What AndroidWorld_SR measures
AndroidWorld_SR is a multimodal benchmark that evaluates large language models on multimodal, reasoning, general, and agents tasks. LLM Stats tracks 8 models on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.5, with the leader reaching 0.7.
Compare leaders on the best AI for multimodal, best AI for reasoning, best AI for general and best AI for agents leaderboards.
Publication
- Paper
- AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents
- Authors
- Christopher Rawles, Sarah Clinckemaillie, Yifan Chang, Jonathan Waltz, and 11 others
- Published
- arXiv
- 2405.14573
Abstract
Autonomous agents that execute human tasks by controlling computers can enhance human productivity and application accessibility. However, progress in this field will be driven by realistic and reproducible benchmarks. We present AndroidWorld, a fully functional Android environment that provides reward signals for 116 programmatic tasks across 20 real-world Android apps. Unlike existing interactive environments, which provide a static test set, AndroidWorld dynamically constructs tasks that are parameterized and expressed in natural language in unlimited ways, thus enabling testing on a much larger and more realistic suite of tasks. To ensure reproducibility, each task includes dedicated initialization, success-checking, and tear-down logic, which modifies and inspects the device's system state. We experiment with baseline agents to test AndroidWorld and provide initial results on the benchmark. Our best agent can complete 30.6% of AndroidWorld's tasks, leaving ample room for future work. Furthermore, we adapt a popular desktop web agent to work on Android, which we find to be less effective on mobile, suggesting future research is needed to achieve universal, cross-platform agents. Finally, we also conduct a robustness analysis, showing that task variations can significantly affect agent performance, demonstrating that without such testing, agent performance metrics may not fully reflect practical challenges. AndroidWorld and the experiments in this paper are available at github.com/google-research/android_world.
Qwen3.5-35B-A3B leads with 71.1%, followed by Qwen3.5-122B-A10B at 66.4% and Qwen3.5-27B at 64.2%.
Progress Over Time
Interactive timeline showing model performance evolution on AndroidWorld_SR
AndroidWorld_SR Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Alibaba Cloud / Qwen Team | 35B | 262K | $0.25 / $2.00 | ||
| 2 | Alibaba Cloud / Qwen Team | 122B | 262K | $0.40 / $3.20 | ||
| 3 | Alibaba Cloud / Qwen Team | 27B | 262K | $0.30 / $2.40 | ||
| 4 | Alibaba Cloud / Qwen Team | 33B | — | — | ||
| 4 | Alibaba Cloud / Qwen Team | 236B | 262K | $0.30 / $1.50 | ||
| 6 | Alibaba Cloud / Qwen Team | 72B | — | — | ||
| 7 | Alibaba Cloud / Qwen Team | 8B | — | — | ||
| 8 | Alibaba Cloud / Qwen Team | 34B | — | — |
FAQ
Common questions about AndroidWorld_SR.
More evaluations to explore
Related benchmarks in the same category
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions