Skip to content

Qwen-Image on M3 Ultra 96 GB: VRAM, speed and fit

Qwen-Image at 16-bit runs on the M3 Ultra 96 GB, about 21 min per image.

Image generation · 1328 × 1328

Runs comfortably · on M3 Ultra 96 GB · 16-bit (BF16)

GPU memory
73.6 GB
of 96.0 GB
Time
21 min
per image

Where the memory goes

Model
45.8 GB
Working
3.0 GB
Overhead
819 MB
OS
24.0 GB
Free
22.4 GB
Memory
73.6 GB of 96.0 GB
Output
1328 × 1328 · 50 steps
Time
21 min per image
Running cost
$5.32/mo at home · $72/mo on a cloud A40

Validation: Not validated

What Qwen-Image on M3 Ultra 96 GB was checked against
CheckTargetResultStatus
Fits in GPU memory72.0 GB usable49.6 GB neededPass
Headroom5% free22.4 GB free (31%)Pass
Time per imageunder 30 s21 minFails

Targets: an image in under 30 s, and 5% of memory left free. Calculated from published specs, not benchmarked.

The setup

Model
Qwen-Image
Parameters
20.4B
Precision
16-bit (BF16)
Output
1328 × 1328
Steps
50
Hardware
M3 Ultra 96 GB
GPU can use
72.0 GB of 96.0 GB

Running cost

At home
$5.32/mo
243 W at the wall, $0.02 an image
Renting instead
$72/mo
A40 at $0.59/h, $0.04 an image

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 22, 2026.

Where to get it

Qwen-Image on similar machines

All machines: Qwen-Image VRAM requirements. Everything that runs here: what can the M3 Ultra 96 GB run?

Other models that run on M3 Ultra 96 GB

Questions

Can Qwen-Image run on M3 Ultra 96 GB?

Qwen-Image at 16-bit runs on the M3 Ultra 96 GB, about 21 min per image.

Report VR-1HWK5RR, calculated Nov 22, 2026 from published specs, not benchmarked.