Qwen Image image benchmarks on AMD GPUs

67 reported configurations · 12 sources · 71 measurements

Documented results

24 selected configurations

Selected across sources, not a performance ranking. Match model weights, resolution, steps, software and timing scope before comparing.

Source	Model	GPUs	Output	Measurements
View ReportComfyUI	Qwen-Image2.1Q8_0Model reference ↗Text → image	Radeon RX 9070 XT1 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS Windows11 Pro build26200 Parallelization Not reported Memory 16 GB per GPU Reported Not reported Full record →	1024 × 1024Batch 120 stepsBatch 1	294.7 s/imageTime / image · reported 15.88 GiBPeak VRAM · reportedPeak allocated, as reported; preserveGiB unit. Whole workflow · Cold
View ReportComfyUI	Qwen-Image-2512 + Lightning LoRAFP8 E4M3FNTested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS Linux (Fedora Rawhide container) Parallelization Single GPU Memory 32 GB per GPU Reported 2026-01-02 Full record →	1328 × 1328Batch 14 stepsBatch 1	33.444 sTime / image · reported Whole workflow · Cold
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.2 exact Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	147.428 sTime / image · median Whole workflow · Cold · median
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.2 exact Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	147.829 sTime / image · reported Whole workflow · Cold
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.1 Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	179.319 sTime / image · median Whole workflow · Cold · median
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.0 Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	179.702 sTime / image · medianMedian of3original process samples; cross-checked againstBENCHMARKS.md. Whole workflow · Cold · median
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.2 wave64 exact attention Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	145.255 sTime / image · median Whole workflow · Cold · median
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.1 Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	177.223 sTime / image · median Whole workflow · Cold · median
View ReportComfyUI	Qwen-Image-Edit-2509 + Lightning LoRAQ4_K_MTested weights ↗Image editing	Radeon AI PRO R97001 GPU Configuration Engine ComfyUI / ComfyUI-GGUF 0.3.64 Backend ROCm / HIP OS Aurora Linux42 (Fedora Kinoite); Ubuntu22.04 container Parallelization Single GPU Memory 32 GB per GPU Reported 2025-10-11 Full record →	Resolution not reportedSteps not reportedBatch not reported	52.17 sReported time · mean Warm · Timing details missing · mean
View ReportComfyUI	Qwen-Image-2512FP8Model reference ↗Text → image	Radeon RX 9060 XT1 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS CachyOS (Arch Linux), kernel 6.19 Parallelization Single GPU Memory 16 GB per GPU Reported Not reported Full record →	1328 × 132850 stepsBatch not reported	415.93 sReported time · reported Timing details missing
View ReportDiffusion workflow	Qwen-Image-2512BF16Model reference ↗Text → image	Radeon RX 7800 XT1 GPU Configuration Engine Gridline Backend Not reported OS Not reported Parallelization Single GPU Memory 16 GB per GPU Reported Not reported Full record →	1664 × 928Batch 150 stepsBatch 1	2,363.6 sReported time · reported Timing details missing
View ReportComfyUI	Qwen-ImageFP8Model reference ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS Not reported Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	1328 × 132850 stepsBatch not reported	135.59 sReported time · reported Timing details missing
View ReportComfyUI	Qwen-Image2.1Q8_0Model reference ↗Text → image	Radeon RX 9070 XT1 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS Windows11 Pro build26200 Parallelization Not reported Memory 16 GB per GPU Reported Not reported Full record →	1024 × 1024Batch 120 stepsBatch 1	40.4 s/imageTime / image · reported 15.88 GiBPeak VRAM · reportedPeak allocated, as reported; preserveGiB unit. Whole workflow · Warm
View ReportComfyUI	Qwen-Image-2512 + Lightning LoRAFP8 E4M3FNTested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS Linux (Fedora Rawhide container) Parallelization Single GPU Memory 32 GB per GPU Reported 2026-01-02 Full record →	1328 × 1328Batch 14 stepsBatch 1	12.355 sTime / image · reported Whole workflow · Warm
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.2 exact Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	136.285 sTime / image · median Whole workflow · Warm · median
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.2 exact Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	133.745 sTime / image · reported Whole workflow · Warm
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.1 Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	165.141 sTime / image · median Whole workflow · Warm · median
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.0 Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	168.101 sTime / image · medianMedian of3original process samples; cross-checked againstBENCHMARKS.md. Whole workflow · Warm · median
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.2 wave64 exact attention Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	133.558 sTime / image · median Whole workflow · Warm · median
View ReportPaiton	Qwen-Image-2.1Balanced MXFP4 / BF16Tested weights ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine Paiton 1.0.1 Backend ROCm / HIP OS Linux Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	2048 × 2048Batch 140 stepsBatch 1	165.953 sTime / image · median Whole workflow · Warm · median
View ReportComfyUI	Qwen-Image-Edit-2509 + Lightning LoRAQ4_K_MTested weights ↗Image editing	Radeon AI PRO R97001 GPU Configuration Engine ComfyUI / ComfyUI-GGUF 0.3.64 Backend ROCm / HIP OS Aurora Linux42 (Fedora Kinoite); Ubuntu22.04 container Parallelization Single GPU Memory 32 GB per GPU Reported 2025-10-11 Full record →	Resolution not reportedSteps not reportedBatch not reported	48.628 sReported time · mean Warm · Timing details missing · mean
View ReportComfyUI	Qwen-Image-2512FP8Model reference ↗Text → image	Radeon RX 9060 XT1 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS CachyOS (Arch Linux), kernel 6.19 Parallelization Single GPU Memory 16 GB per GPU Reported Not reported Full record →	1328 × 132850 stepsBatch not reported	395.19 sReported time · reported Timing details missing
View ReportComfyUI	Qwen-ImageFP8Model reference ↗Text → image	Radeon AI PRO R97001 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS Not reported Parallelization Single GPU Memory 32 GB per GPU Reported Not reported Full record →	1328 × 132850 stepsBatch not reported	103.78 sReported time · reported Timing details missing
View ReportComfyUI	Qwen-Image2.1INT8 ConvRotTested weights ↗Text → image	Radeon RX 9070 XT1 GPU Configuration Engine ComfyUI Backend ROCm / HIP OS Windows11 Pro build26200 Parallelization Not reported Memory 16 GB per GPU Reported Not reported Full record →	1024 × 1024Batch 120 stepsBatch 1	155.9 s/imageTime / image · reported 15.82 GiBPeak VRAM · reportedPeak allocated, as reported; preserveGiB unit. Whole workflow · Cold

Explore all 67 configurations →

All image benchmarks · Methodology and source review