Skip to content

Qwen-Image 2.1 VRAM requirements

How much VRAM Qwen-Image 2.1 needs
8-bit: 8.7 GB4-bit GGUF: 5.8 GBAt 2048 × 2048
Type
Image model
Parameters
7.1B
Text encoder
Qwen3-VL-8B-Instruct
Released
Sep 2026
License
Qwen Research License

How much VRAM Qwen-Image 2.1 needs

Memory Qwen-Image 2.1 needs at each precision
PrecisionWeightsPeak at 2048 × 2048
16-bit BF1613.3 GB15.1 GB
8-bit FP8 · Q8_06.9 GB8.7 GB
6-bit GGUF Q6_K5.5 GB7.3 GB
5-bit GGUF Q5_K_M4.7 GB6.6 GB
4-bit GGUF Q4_K_M · NVFP44.0 GB5.8 GB

The Qwen3-VL-8B-Instruct text encoder (8.2 GB at FP8) runs first and then waits in system RAM.

GPUs and Macs that run Qwen-Image 2.1

How Qwen-Image 2.1 runs on common GPUs, Macs and AI boxes
MachineBest fitPer imageReport
RTX 4060Too bigPDF: Qwen-Image 2.1 on RTX 4060 report
RTX 5060Too bigPDF: Qwen-Image 2.1 on RTX 5060 report
RTX 3070Too bigPDF: Qwen-Image 2.1 on RTX 3070 report
RTX 3080 10 GB8-bit4.2 minPDF: Qwen-Image 2.1 on RTX 3080 10 GB report
RTX 3060 12 GB8-bit9.7 minPDF: Qwen-Image 2.1 on RTX 3060 12 GB report
Arc B5808-bit5.1 minPDF: Qwen-Image 2.1 on Arc B580 report
RTX 40708-bit4.3 minPDF: Qwen-Image 2.1 on RTX 4070 report
RTX 4070 Super8-bit3.5 minPDF: Qwen-Image 2.1 on RTX 4070 Super report
RTX 50708-bit4.0 minPDF: Qwen-Image 2.1 on RTX 5070 report
RTX 4060 Ti 16 GB16-bit5.6 minPDF: Qwen-Image 2.1 on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GB16-bit5.2 minPDF: Qwen-Image 2.1 on RTX 5060 Ti 16 GB report
Arc A770 16 GB16-bit4.5 minPDF: Qwen-Image 2.1 on Arc A770 16 GB report
Radeon RX 9070 XT16-bit2.2 minPDF: Qwen-Image 2.1 on Radeon RX 9070 XT report
RTX 4070 Ti Super16-bit2.8 minPDF: Qwen-Image 2.1 on RTX 4070 Ti Super report
RTX 4080 Super16-bit2.4 minPDF: Qwen-Image 2.1 on RTX 4080 Super report
RTX 5070 Ti16-bit2.8 minPDF: Qwen-Image 2.1 on RTX 5070 Ti report
RTX 508016-bit2.2 minPDF: Qwen-Image 2.1 on RTX 5080 report
Radeon RX 7900 XTX16-bit3.5 minPDF: Qwen-Image 2.1 on Radeon RX 7900 XTX report
RTX 309016-bit3.5 minPDF: Qwen-Image 2.1 on RTX 3090 report
RTX 409016-bit1.5 minPDF: Qwen-Image 2.1 on RTX 4090 report
Radeon AI PRO R970016-bit2.2 minPDF: Qwen-Image 2.1 on Radeon AI PRO R9700 report
RTX 509016-bit72 sPDF: Qwen-Image 2.1 on RTX 5090 report
RTX PRO 5000 Blackwell 48 GB16-bit58 sPDF: Qwen-Image 2.1 on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada16-bit42 sPDF: Qwen-Image 2.1 on RTX 6000 Ada report
L40S16-bit42 sPDF: Qwen-Image 2.1 on L40S report
RTX PRO 6000 Blackwell16-bit31 sPDF: Qwen-Image 2.1 on RTX PRO 6000 Blackwell report
Mac mini M4 16 GBToo bigPDF: Qwen-Image 2.1 on Mac mini M4 16 GB report
Mac mini M4 24 GB4-bit74 minPDF: Qwen-Image 2.1 on Mac mini M4 24 GB report
Mac mini M4 32 GB8-bit64 minPDF: Qwen-Image 2.1 on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GB4-bit37 minPDF: Qwen-Image 2.1 on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB16-bit32 minPDF: Qwen-Image 2.1 on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB16-bit32 minPDF: Qwen-Image 2.1 on Mac mini M4 Pro 64 GB report
M4 Max 128 GB16-bit16 minPDF: Qwen-Image 2.1 on M4 Max 128 GB report
M5 Max 128 GB16-bit4.7 minPDF: Qwen-Image 2.1 on M5 Max 128 GB report
M3 Ultra 96 GB16-bit9.1 minPDF: Qwen-Image 2.1 on M3 Ultra 96 GB report
M3 Ultra 256 GB16-bit9.1 minPDF: Qwen-Image 2.1 on M3 Ultra 256 GB report
M3 Ultra 512 GB16-bit9.1 minPDF: Qwen-Image 2.1 on M3 Ultra 512 GB report
M5 Ultra 512 GB16-bit2.8 minPDF: Qwen-Image 2.1 on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB16-bit7.1 minPDF: Qwen-Image 2.1 on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB16-bit7.1 minPDF: Qwen-Image 2.1 on Ryzen AI Max+ 395 128 GB report
RTX Spark16-bit4.0 minPDF: Qwen-Image 2.1 on RTX Spark report
DGX Spark16-bit4.0 minPDF: Qwen-Image 2.1 on DGX Spark report
DGX Station GB30016-bit6.9 sPDF: Qwen-Image 2.1 on DGX Station GB300 report
A100 80 GB SXM16-bit49 sPDF: Qwen-Image 2.1 on A100 80 GB SXM report
H100 SXM16-bit16 sPDF: Qwen-Image 2.1 on H100 SXM report

2048 × 2048, 40 steps. Times are rough, plus or minus 30%.

Where to get it

Questions

How much VRAM does Qwen-Image 2.1 need?

Qwen-Image 2.1 needs about 8.7 GB at 8-bit, or 5.8 GB as a 4-bit GGUF, at its default output size.

What is the minimum VRAM for Qwen-Image 2.1?

The minimum is about 5.8 GB of VRAM, as a 4-bit file at its default output size, with the text encoder parked in system RAM.

How big is Qwen-Image 2.1?

The 7.1B denoiser is 13.3 GB at 16-bit and 6.9 GB at 8-bit; the Qwen3-VL-8B-Instruct text encoder comes on top.

Can Qwen-Image 2.1 run on an RTX 4090?

Yes. At 16-bit it takes about 1.5 min per image.

More models