Benchmark
Blog
Join Cloud Waitlist
minitap.ai
Benchmark
Blog
Join Cloud Waitlist
Menu
Last Update
Sep 15th 2025
AndroidWorld Benchmark: Our Evolution
A look at our journey to the top of the androidWorld benchmark, and our new record-breaking score.
#1
Ranking
84.48%
Current ScoreRanki
116
Tasks Evaluated
Benchmark Comparison
See how we compare against the competition.
View Complete AndroidWorld Leaderboard
Complete Task Analysis
Detailed breakdown of all 116 benchmark tasks with transparent trace data.
Our Journey to the Top
Launch
Minitap Launch
June 15th, 2025
Launched minitap mobile-use with the ambitious goal of reaching #1 on the leaderboard within 2 weeks.
Read "The builders of mobile AI"
68.1%
Public Announcement
July 1st, 2025
Public announcement of our 68.1% score, establishing our presence on the AndroidWorld leaderboard.
Read "Comprehensive evaluation of mobile AI agents"
74.14%
Continuous Improvement
August 20th, 2025
Evolution to 74.14%, strengthening our leading position on the AndroidWorld benchmark.
Read "The builders of mobile AI"
77.6%
Industry Record
September 11th, 2025
New record at 77.6% thanks to Cortex meta-reasoning system improvements and tool validation enhancements, solidifying our position as the undisputed leader.
Read "Back to State-of-the-Art: 77.59%"
Open Source Release
Explore our mobile AI agents benchmark, contribute to the codebase, and help advance mobile automation research. Our work is open source, enabling the community to build upon it and advance mobile AI agent capabilities together.
View Repository
1.6k
Github
minitap.ai
Benchmark
Blog