[ 50 models ] [ 99 GPUs ] [ 29 measurements on a real 16 GB card ]

Will FLUX.1 [dev]
run on your GPU?

FLUX, Qwen-Image, Krea 2, Wan, LTX and 45 more. Which file to download, how much VRAM it really needs, and what your card can handle. Sizes from the real files, speed measured in ComfyUI.

I have Your GPU
and want to run AI model

…

Loading the answer.

[ FLUX.1 schnell ][ flux1-schnell-Q8_0.gguf ][ 9.4 s ][ 15.1 GB peak ][ FLUX.1 dev ][ flux1-dev-Q8_0.gguf ][ 40.6 s ][ 15.1 GB peak ][ FLUX.1 dev ][ flux1-dev-Q4_K_S.gguf ][ 47.3 s ][ 8.8 GB peak ][ FLUX.1 dev ][ flux1-dev-fp8.safetensors ][ 36.4 s ][ 14.8 GB peak ][ Wan 2.2 5B ][ wan2.2_ti2v_5B_fp16.safetensors ][ 8.6 min ][ 11.0 GB peak ][ Z-Image Turbo ][ z_image_turbo_bf16.safetensors ][ 11.0 s ][ 14.6 GB peak ][ Wan 2.2 T2V ][ low_noise_14B_fp8_scaled.safetenso ][ 19 min ][ 13.3 GB peak ][ Wan 2.2 T2V ][ LowNoise-Q5_K_M.gguf ][ 26 min ][ 14.0 GB peak ][ Wan 2.2 T2V ][ LowNoise-Q8_0.gguf ][ 25 min ][ 13.1 GB peak ][ FLUX.2 klein 4B ][ flux-2-klein-4b-fp8.safetensors ][ 2.9 s ][ 12.2 GB peak ][ FLUX.2 klein 4B ][ flux-2-klein-4b.safetensors ][ 3.9 s ][ 14.1 GB peak ][ FLUX.2 klein 4B ][ flux-2-klein-4b-Q8_0.gguf ][ 4.3 s ][ 12.4 GB peak ][ FLUX.2 klein 9B ][ flux-2-klein-9b-Q8_0.gguf ][ 8.5 s ][ 13.5 GB peak ][ FLUX.2 klein 9B ][ flux-2-klein-9b-Q4_K_M.gguf ][ 9.8 s ][ 13.5 GB peak ][ Qwen-Image 2.1 ][ qwen_image_2.1_int8_convrot.safete ][ 17.2 s ][ 14.3 GB peak ][ Qwen-Image 2.1 ][ qwen_image_2.1_Q8_0.gguf ][ 43.6 s ][ 14.8 GB peak ][ Qwen-Image 2.1 ][ qwen_image_2.1_bf16.safetensors ][ 39.7 s ][ 15.5 GB peak ][ Krea 2 ][ krea2_turbo_int8_convrot.safetenso ][ 10.9 s ][ 15.3 GB peak ][ Krea 2 ][ krea2_turbo_fp8_scaled.safetensors ][ 17.7 s ][ 14.9 GB peak ][ Ming-Image ][ ming_image_0.1_design_int8_convrot ][ 7.4 s ][ 13.6 GB peak ][ Ming-Image ][ ming_image_0.1_design_int8_convrot ][ 44.9 s ][ 14.4 GB peak ][ Ming-Image ][ ming_image_0.1_design_bf16.safeten ][ 17.4 s ][ 15.3 GB peak ][ FLUX.1 Kontext ][ flux1-dev-kontext_fp8_scaled.safet ][ 53.4 s ][ 15.0 GB peak ][ FLUX.1 Kontext ][ flux1-kontext-dev-Q8_0.gguf ][ 82.1 s ][ 15.5 GB peak ][ Qwen-Image-Edit ][ qwen-image-edit-2511-Q4_K_M.gguf ][ 27.2 s ][ 15.9 GB peak ][ Qwen-Image ][ qwen-image-Q4_K_M.gguf ][ 123.3 s ][ 15.7 GB peak ][ Chroma1-HD ][ Chroma1-HD-fp8mixed.safetensors ][ 54.4 s ][ 15.4 GB peak ][ Z-Image ][ z_image_bf16.safetensors ][ 78.0 s ][ 15.5 GB peak ][ HunyuanVideo 1.5 ][ hunyuanvideo1.5_480p_t2v_cfg_disti ][ 11 min ][ 14.3 GB peak ]

50image & video models

99desktop & laptop GPUs

851model files, sized to the byte

16GB card in my PC, where every timing here was run

01Measured

Real outputs. Real seconds.

Every image came out of my RTX 5060 Ti 16 GB, with a workflow you can download. All workflows →

FLUX.1 [dev] sample made on an RTX 5060 Ti 16 GB

FLUX.1 [dev]FP8 · 1024²36.4s

Z-Image Turbo sample made on an RTX 5060 Ti 16 GB

Z-Image Turbo16-bit · 1024²11.0s

Krea 2 Turbo sample made on an RTX 5060 Ti 16 GB

Krea 2 TurboINT8 · 1024²10.9s

Qwen-Image 2.1 sample made on an RTX 5060 Ti 16 GB

Qwen-Image 2.1INT8 · 1024²17.2s

Wan 2.2 T2V 14B sample made on an RTX 5060 Ti 16 GB

Wan 2.2 T2V 14BFP8 · 832×480 · video19min

FLUX.2 [klein] 4B sample made on an RTX 5060 Ti 16 GB

FLUX.2 [klein] 4BFP8 · 1024²2.9s

Ming-Image Design sample made on an RTX 5060 Ti 16 GB

Ming-Image DesignINT8 · 1024²7.4s

Wan 2.2 TI2V 5B sample made on an RTX 5060 Ti 16 GB

Wan 2.2 TI2V 5B16-bit · 1280×704 · video8.6min

FLUX.1 Kontext sample made on an RTX 5060 Ti 16 GB

FLUX.1 KontextFP8 · photo edit53.4s

FLUX.2 [klein] 9B sample made on an RTX 5060 Ti 16 GB

FLUX.2 [klein] 9BQ8_0 · 1024²8.5s

02By VRAM

Start from the card you have.

Each dot is one of the 50 models. All VRAM sizes →

VRAMTIER 018GB

RTX 4060 · 5060 · 3070
RTX 3060 Ti · 2080 Super
RX 7600 · 9060 XT 8 GB

Runs well 8Runs 7Tight 8Offload 27 VRAMTIER 0212GB

RTX 5070 · 4070 Super
RTX 3060 12 GB · 4070 Ti
Arc B580 · RX 7700 XT

Runs well 18Runs 8Tight 12Offload 12 VRAM● TESTED HERE16GB

RTX 5060 Ti · 4060 Ti 16 GB
RTX 5080 · 4080 · 5070 Ti
RX 9070 XT · 7800 XT

Runs well 26Runs 11Tight 11Offload 2 VRAMTIER 0424GB

RTX 4090 · 3090
RX 7900 XTX
Arc Pro B60

Runs well 44Runs 5Tight 1Offload 0

Runs wellRuns (compressed file)TightOffload onlyNot practical

6 GB 10 GB 20 GB 32 GB

03The memory ruler

Every model on one scale. Drag the red line to your VRAM.

Each dot is one file of a model, placed at the VRAM it needs. Everything left of the line fits.

Your VRAM

16GB

0 of 50 models fit entirely.

VRAM in GB

Model

081624324048

Verdict · file

Image models

Video models

A file of this model (VRAM it needs)Fits in your VRAMBest file that fitsYour VRAM

04Asked most

Plan the setting, the upgrade, the laptop

CalculatorBigger images, longer clips, LoRAs and ControlNetsSet resolution, frames and batch; see which file still fits → UpgradeWhat would a new GPU change?Your card vs the one you want: what newly runs well → TrainingLoRA training: how much VRAM49 figures from the trainers' own docs, with settings → LaptopsLaptop GPUs have less VRAM than their names suggestThe RTX 4090 Laptop has 16 GB, not 24. 32 laptop GPUs → FilesEvery text encoder, VAE and LoRA explainedWhich folder it goes in, which models share it → Compare44 GPU matchups, model by modelRTX 3090 vs 5060 Ti, 7900 XTX vs 4090… →

07No guesses without a label

Where every number comes from

How the numbers work →

calculated

From real file sizes

Every model file is sized to the byte from Hugging Face. Add the model's working memory and a small system reserve, and you get the verdict of what fits entirely.

measured

On my own GPU

I make AI images in ComfyUI on an RTX 5060 Ti 16 GB. A script times each model and logs peak VRAM: 16 models timed so far. See the numbers or get the workflows I used.

reported

From other people

98 real-world results from GitHub, Hugging Face and first-hand blogs, copied as published with a link to each. See them all or send yours.

08Newest first

Recent models

Month by month, with email alerts →

2026-09Qwen-Image 2.17.1B · image · fits from 6 GB 2026-09Ming-Image 0.1 Design6.15B · image · fits from 8 GB 2026-07Mage-Flow (Microsoft)4.12B · image · fits from 6 GB 2026-07Wan Animate 2 (14B)14B · video · fits from 12 GB 2026-07LTX-2.5 (22B)22B · video · fits from 14 GB 2026-07MiniMax H3 (33B)33.1B · video · fits from 22 GB 2026-07MiniMax H3 Prunedpruned 33B · video · fits from 15 GB 2026-06Krea 2 (Turbo)12.8B · image · fits from 8 GB

09Short and practical

Guides

G.01What runs on an RTX 5060 Ti 16 GB in ComfyUI: every model, timedMeasured G.0212 GB or 16 GB VRAM for ComfyUI: what the extra 4 GB gets youBuying G.03Image editing on 16 GB: FLUX.1 Kontext vs Qwen-Image-Edit 2511, timedMeasured G.04Best local image model for 16 GB: 12 models tested side by sideMeasured G.05Can you use it commercially? Licences of local AI modelsLicences G.06Qwen-Image 2.1, Krea 2 and Ming-Image on 16 GB: timedMeasured G.07FLUX.2 klein 4B and 9B on a 16 GB RTX 5060 Ti: timedMeasured G.08Wan 2.2 14B on a 16 GB card: FP8 streaming vs GGUF, timedMeasured G.09FP8 vs GGUF on a 16 GB RTX 5060 Ti: I timed FLUX.1 devMeasured G.10GGUF Q8, Q6, Q5, Q4: which one should you download?Files G.11FP8, GGUF or 16-bit in ComfyUI: what is the difference?Files G.12How much VRAM do you need for AI images and video in 2026?Buying G.13Out of memory in ComfyUI: fixes in the order to try themFixes G.14Text encoders: why T5, UMT5 and Qwen-VL eat your memoryExplainer G.15How much system RAM do you need for ComfyUI?Buying G.16AMD and Intel GPUs for ComfyUI in 2026GPUs G.17Local AI on a laptop: what the GPU name does not tell youGPUs G.1896 GB of "VRAM" on Strix Halo: what fits, and what it costs youGPUs G.19Why video models need so much more memory than image modelsExplainer G.20INT8, NVFP4, MXFP8: the new model file formats of 2026Files

Will FLUX.1 [dev]
run on your GPU?

Real outputs. Real seconds.

Start from the card you have.

Every model on one scale. Drag the red line to your VRAM.

Popular questions

The exact workflow behind every number.

Plan the setting, the upgrade, the laptop

Where every number comes from

From real file sizes

On my own GPU

From other people

Recent models

Guides

Will FLUX.1 [dev]run on your GPU?

Real outputs. Real seconds.

Start from the card you have.

Every model on one scale. Drag the red line to your VRAM.

Popular questions

The exact workflow behind every number.

Plan the setting, the upgrade, the laptop

Where every number comes from

From real file sizes

On my own GPU

From other people

Recent models

Guides

Will FLUX.1 [dev]
run on your GPU?