Personal assistant benchmark: Gemini Spark leads
micro1 scores Instinct, Muse, Grok Bot and Gemini Spark on real users' everyday tasks, with Gemini leading thanks to Google's ecosystem.
If you're using a personal agent like Instinct, Muse, Grok Bot, or Gemini Spark, this benchmark is for you. We're in the early days of personal AI agents and we wanted to build a reliable way to compare them. Instead of running scripted synthetic tests, we replicated how a normal person would actually judge an assistant. The tasks come from everyday life, the testers are real users running them in their own accounts, and every result is scored twice: once on whether the task got done, and once on whether it got done in a way you'd be comfortable with. Gemini Spark leading might be surprising, but there’s a simple reason: many everyday personal agent tasks happen inside Google’s ecosystem—Gmail, Google Flights, Docs, Calendar, etc. When an agent needs to actually take actions across these products, Google has a natural advantage as it can connect to its own products more reliably and accurately than anyone else. This is a living leaderboard and results will update as more users contribute. What you see today is truly just the beginning. If you use one of these agents and want your experience to count (and get paid for it), the link to join is in the post below.
This post quotes another post: x.com/aliansarinik/status/2107497652871155953