Qwen-Image 2.1 on M3 Ultra 96 GB: VRAM, speed and fit
Qwen-Image 2.1 at 16-bit runs on the M3 Ultra 96 GB, about 9.1 min per image.
Image generation · 2048 × 2048
Runs comfortably · on M3 Ultra 96 GB · 16-bit (BF16)
- GPU memory
- 49.2 GB
- of 96.0 GB
- Time
- 9.1 min
- per image
Where the memory goes
- Model
- 21.4 GB
- Working
- 3.0 GB
- Overhead
- 819 MB
- OS
- 24.0 GB
- Free
- 46.8 GB
- Memory
- 49.2 GB of 96.0 GB
- Output
- 2048 × 2048 · 40 steps
- Time
- 9.1 min per image
- Running cost
- $5.32/mo at home · $29/mo on a cloud RTX 2000 Ada
Validation: Not validated
| Check | Target | Result | Status |
|---|---|---|---|
| Fits in GPU memory | 72.0 GB usable | 25.2 GB needed | Pass |
| Headroom | 5% free | 46.8 GB free (65%) | Pass |
| Time per image | under 30 s | 9.1 min | Fails |
Targets: an image in under 30 s, and 5% of memory left free. Calculated from published specs, not benchmarked.
The setup
- Model
- Qwen-Image 2.1
- Parameters
- 7.1B
- Precision
- 16-bit (BF16)
- Output
- 2048 × 2048
- Steps
- 40
- 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, $0.007 an image
- Renting instead
- $29/mo
- RTX 2000 Ada at $0.24/h, $0.02 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 2.1 on similar machines
- Mac mini M4 Pro 48 GBRuns
- Mac mini M4 Pro 64 GBRuns
- M4 Max 128 GBRuns
- M5 Max 128 GBRuns
- M3 Ultra 256 GBRuns
- M3 Ultra 512 GBRuns
- M5 Ultra 512 GBRuns
- Ryzen AI Max+ 395 64 GBRuns
All machines: Qwen-Image 2.1 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 2.1 run on M3 Ultra 96 GB?
Qwen-Image 2.1 at 16-bit runs on the M3 Ultra 96 GB, about 9.1 min per image.
Report VR-0NJ1ZH4, calculated Nov 22, 2026 from published specs, not benchmarked.