multimodal benchmark
MobileWorld
MobileWorld is a benchmark for evaluating multimodal agents on real mobile-device tasks, testing GUI grounding, navigation, and multi-step task completion in mobile environments.
Updated Aug 7, 2026
Published models3
Registry coverage3
MetricScore
EvidenceB
MobileWorld leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AC | 77.8% | 100.0% | 3 | C | |
| 2 | BY | 73.1% | 50.0% | 3 | C | |
| 3 | BY | 70.0% | 0.0% | 3 | C |
Score distribution
Top published rows on the benchmark's original scale.
MobileWorld highlights
The top published results on this benchmark's own scale.
What is MobileWorld?
Definition and scoring fields from the benchmark registry.
MobileWorld is a benchmark for evaluating multimodal agents on real mobile-device tasks, testing GUI grounding, navigation, and multi-step task completion in mobile environments.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- MobileWorld
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- mobileworld|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about MobileWorld.
Which model scores highest on MobileWorld?
Qwen3.8 Max is currently ranked first with 77.8%.
What does MobileWorld measure?
MobileWorld is a benchmark for evaluating multimodal agents on real mobile-device tasks, testing GUI grounding, navigation, and multi-step task completion in mobile environments.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
3 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.