MobileWorld - GUI Only: leaderboard
The GUI-only split of MobileWorld (Tongyi-MAI): 117 smartphone-environment tasks solved end-to-end from screen interaction alone; success rate %, agentic planner+grounder combos excluded.
Metric: Success Rate (%). Source: tongyi-mai.github.io. Saturation forecast: Around February 2027. 36 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K3 | 74.4 | #35 |
| 2 | GPT-5.6 Sol | 70.1 | #22 |
| 3 | Seed 2.0 Pro | 63.2 | #122 |
| 4 | GPT-5.5 | 62.4 | #33 |
| 5 | Gemini 3.1 Pro (Preview) | 58.1 | #64 |
| 6 | Claude Opus 4.7 | 56.4 | #51 |
| 7 | Kimi K2.6 | 55.6 | #113 |
| 8 | Seed 1.8 | 52.1 | #143 |
| 9 | Gemini 3 Pro | 51.3 | #82 |
| 10 | Kimi K2.5 | 49.6 | #152 |
| 11 | Claude Sonnet 4.5 | 47.8 | #148 |
| 12 | Claude Opus 4.6 | 44.5 | #67 |
| 13 | Qwen 3.5 397B A17B | 42.7 | #154 |
| 14 | Qwen 3.5 122B A10B | 35 | #184 |
| 15 | Qwen 3 VL 235B A22B | 12.8 | #251 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mobileworld-gui-only · How It Works · Data refreshed daily, snapshot 2026-11-05.