Answer · updated 2026-10-18
Which AI coding model or tool do engineers actually prefer right now?
Sensus measures it weekly: in spaces such as Harness Engineers, Inference Engineers, Claude Builders and XAI Fanatics, members answer questions like which model they switched to this month and which bottleneck still blocks shipped agents, the shares are counted, and the ranking and its week-over-week move are shown in the app. It is the closest thing to a live poll of working engineers on Claude, Codex, Cursor, Grok and the rest.
What Sensus shows
- Weekly rankings per room: the selections that lead and by how much.
- Highlights: the room's questions of the week, with the leading answer and its share.
- Readings on the themes engineers keep voting on: performance on real projects, reliability of the full workflow, cost per completed task, freedom to switch models.
- A reputable AI read that answers a question only from these votes, with what changed this week.
Why it is different from a benchmark or a leaderboard
A benchmark scores a model on a task set. Sensus counts what working engineers choose for their own work, week by week, inside a verified room, so it tracks preference and switching rather than scores.
Join the Sensus waitlistSensus is an invite-only private beta on iOS. The waitlist is the way in; members can also send an invite link.
Questions people ask
Can I see the numbers without joining?
The readings live in the app for members. The waitlist is the way in; the beta is invite-only.
Which rooms vote on models?
Harness Engineers, Inference Engineers, Eval Engineers, Agentic Engineers, Frontier Labs, Claude Builders and XAI Fanatics, among others.
How often does it update?
Votes count continuously; highlights and rankings are weekly, and readings carry a week-over-week move.