multimodal benchmark
AITZ_EM Leaderboard
AITZ_EM is a multimodal benchmark for evaluating autonomous GUI agents on smartphones using 18,643 screen-action pairs from over 70 Android apps.
Updated Sep 5, 2026
Models3
Model coverage3
MetricScore
EvidenceB
AITZ_EM Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAC | Score83.20% | Percentile100.00% | Participants3 | EvidenceC | Evaluated |
| Rank02 | ModelAC | Score83.10% | Percentile50.00% | Participants3 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score81.90% | Percentile0.00% | Participants3 | EvidenceC | Evaluated |
AITZ_EM Highlights
The leading models and scores on this benchmark.
AITZ_EM Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for AITZ_EM
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is AITZ_EM?
What AITZ_EM measures and how its scores work.
AITZ_EM is the Android-In-The-Zoo (AitZ) benchmark, with screen-action pairs and chain-of-action-thought annotations for natural language-triggered smartphone task completion.
It measures autonomous GUI-agent performance in smartphone task completion, including screen perception and action decision-making, using Score reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- AITZ_EM
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about AITZ_EM.
Which model scores highest on AITZ_EM?
Qwen2.5 VL 72B Instruct is currently ranked first with 83.20%.
What are the top three models on AITZ_EM?
The current leaders are Qwen2.5 VL 72B Instruct (83.20%), Qwen2.5 VL 32B Instruct (83.10%), and Qwen2.5 VL 7B Instruct (81.90%).
Which AITZ_EM model has the lowest official input price?
Qwen2.5 VL 7B Instruct has the lowest matched official input price at $0.35 input / $1.1 output per 1M tokens.
Which models are fastest among AITZ_EM results?
No matched runtime record is currently available.
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.