MobileForge - Navigation: leaderboard
MobileForge (Harbin Institute of Technology and others, 2026) evaluates project-level mobile app generation: from the screenshots of 29 real apps, an agent in a fixed harness builds a React, TypeScript and Tailwind codebase. This board is the percent of 701 page-existence, tab and parent-child navigation tests the generated apps pass; higher is better.
Metric: Navigation pass rate (%; share of 701 page-existence, tab-navigation and parent-child navigation test specifications passed by the generated app under state-isolated execution; 29 real mobile apps given as 309 screenshots with page-relationship annotations, each turned into a React, TypeScript and Tailwind project by one tool-using agent in a fixed harness (up to 50 iterations, build errors fed back); one run per app; higher is better). Source: arxiv.org. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 92 | #102 |
| 2 | GPT-5 | 89 | #167 |
| 3 | Claude Haiku 4.5 | 85 | #568 |
| 4 | Gemini 2.5 Pro | 82 | #266 |
| 5 | Gemini 2.5 Flash | 62 | #453 |
| 6 | GPT-5 Mini | 58 | #348 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=mobileforge-navigation · How It Works · Data refreshed daily, snapshot 2026-11-13.