StillOn

5 apps · 15 matched runs

We closed the lid 15 times. StillOn used the least power every time.

Five apps. Three samples each. Same unplugged M4 MacBook Pro.

Closed-lid power draw

Lower is better
  • StillOn1.76 W
  • Clamshell2.90 W
  • pmset0% brightness2.98 W
  • Vorssaint0% brightness2.99 W
  • Amphetamine0% brightness3.03 W
  • Vorssaint100% brightness9.73 W
  • pmset100% brightness9.74 W
  • Amphetamine100% brightness9.80 W
Mean of 3 matched idle runs per app and brightness level.

The other apps leave your screen on.

Turn the brightness down and Amphetamine, Vorssaint and pmset drop from about 9.8 W to 3.0 W. That gap is the backlight, still running behind a shut lid. StillOn sat near 1.8 W either way, because macOS reported its display asleep in every sample we took.

Clamshell held near 2.9 W at both brightness levels. macOS still called its display awake, and because brightness barely moved its number, we cannot pin that gap on the backlight alone.

Power draw over the two brightness phasesA 31-second rolling mean of the three runs for each product, aligned to show 100 percent brightness for the first two minutes and 0 percent brightness for the final two minutes.0 W2 W4 W6 W8 W10 W0 min1 min2 min3 min4 min100% brightness0% brightness
31-second rolling mean of 1 Hz samples from 3 runs per app. Counterbalanced runs are aligned by brightness phase.

Every number we recorded.

The idle estimate divides the same 55.1 Wh of battery by each app's measured draw. Running actual agents will drain it faster than sitting idle does.

MeasureStillOnClamshellVorssaintpmsetAmphetamine
100% brightness1.80 W2.93 W9.73 W9.74 W9.80 W
0% brightness1.71 W2.86 W2.99 W2.98 W3.03 W
Mean power1.76 W2.90 W6.35 W6.35 W6.40 W
Built-in displayasleepawakeawakeawakeawake
Idle estimate on 80% battery31.3 hours19.0 hours8.7 hours8.7 hours8.6 hours

Check the numbers.

Every accepted run, display observation and 1 Hz power sample is published in full. Recalculate the result or inspect the source measurements yourself.

Independently run. Unsponsored.

StillOn commissioned, designed and ran this first-party study without sponsorship, payment or input from any competitor or third party. It is independent in funding and execution, not a third-party endorsement.

One Mac, one narrow question.

This idle screening used internal telemetry on one Mac. It does not establish AI-task energy, thermal safety or the same result on every Mac.

Machine
Mac16,1, Apple M4, 24 GB
System
macOS 27.0, build 26A5425a
Protocol
Battery, Wi-Fi off, lid closed, no external display
Sampling
AppleSMC.PSTR at 1 Hz, three valid runs per app