Benchmark

AndroidWorld_SR Benchmark

What is AndroidWorld_SR?

AndroidWorld Success Rate (SR) benchmark - A dynamic benchmarking environment for autonomous agents operating on Android devices. Evaluates agents on 116…

Released
2024-05-23
Modality
multimodal
Evaluates
multimodal, reasoning, general, agents
Openness
unknown
Importer
LLM Stats
Reported scores
8

Source provenance

AndroidWorld_SR results and reported scores

LLM Stats

ModelOrganizationReported valueReportedEvidence
Qwen2.5 VL 32B InstructQwen0.222025-02-28evidence
Qwen2.5 VL 72B InstructQwen0.352025-01-26evidence
Qwen2.5 VL 7B InstructQwen0.2552025-01-26evidence
Qwen3 VL 235B A22B InstructQwen0.6372025-09-22evidence
Qwen3 VL 32B ThinkingQwen0.6372025-09-22evidence
Qwen3.5-122B-A10BQwen0.6642026-02-24evidence
Qwen3.5-27BQwen0.6422026-02-24evidence
Qwen3.5-35B-A3BQwen0.7112026-02-24evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.

Which sources cite AndroidWorld_SR?

Related multimodal, reasoning, general, agents benchmarks