Stable Diffusion Versions

 info from wiki https://en.wikipedia.org/wiki/Stable_Diffusion

Version numberRelease dateParametersNotes
1.1, 1.2, 1.3, 1.4August 2022
All released by CompVis. There is no "version 1.0". 1.1 gave rise to 1.2, and 1.2 gave rise to both 1.3 and 1.4.
1.5October 2022983MInitialized with the weights of 1.2, not 1.4. Released by RunwayML.
2.0November 2022
Retrained from scratch on a filtered dataset.
2.1December 2022
Initialized with the weights of 2.0.
XL 1.0July 20233.5BThe XL 1.0 base model has 3.5 billion parameters, making it around 3.5x larger than previous versions.
XL TurboNovember 2023
Distilled from XL 1.0 to run in fewer diffusion steps.
3.0February 2024 (early preview)800M to 8BA family of models.
3.5October 20242.5B to 8BA family of models with Large (8 billion parameters), Large Turbo (distilled from SD 3.5 Large), and Medium (2.5 billion parameters).
cascadeDecember 2024

Models Summary for IPEX SD.Next image generation

ModelResolutionStepsCFGSamplerTimetestcomments
FLUX.1-dev512x512
768x768		20
100		2-3
0-9+		Euler FlowMatch
DPM2++ SDE FlowMatch
2m 30.14s
11m 30.82s		Test 5 - FLUX.1-dev - steps and guidance
Test 6 - FLUX.1-dev - Sampling methods		defaults 1024x1024, CFG 3.5, Steps 50
Need to sign in and agree to the license
SANA1.5_4.8B_1024px768x768
1024x1024		20+*Default, Euler FlowMatch
DPM2++ 2M FlowMatch
2m 19.11s Test 23 - SANA 1.5 4.8B 1024px - Stepshttps://nvlabs.github.io/Sana/Sana-1.5/
defaults 1024x1024, CFG 4.5, Steps 20

Lumina-Image-2.0

512x512
1280x1280

20
100

2-9+Default1m 29.33s
21m 50.30s		Test 8 - Lumina-Image-2.0 - steps and guidancedefaults 1024x1024, CFG 4.0, Steps 50
Chroma768x768

20
50

2-4Euler FlowMatch
DPM2++ SDE FlowMatch
21m 19.45s		Test 19 - Chroma - steps and guidancedefaults 1024x1024, CFG 3.5, Steps 50
shuttle-3.1-aesthetic768x768

16+

*DPM2++ 2M FlowMatch3m 39.65sTest 21 - shuttle-3.1-aesthetic - steps and samplersdefaults 1024x1024, CFG 3.5, Steps 4
OmniGen v1512x51220+2+Default2m 11.91sTest 10 - OmniGen-v1 - steps and guidancedefaults 1024x1024, CFG 2.5, Steps 50
OmniGen v21024x102432+*Default24m 51.37sTest 22 - OmniGen-v2 - steps and guidance
HiDream-I1-Full1024x1024

20

50+

2-6

0+

Default

15m 45.30s 

38m 49.60s 

Test 12 - HiDream I1 Full - steps and guidancedefaults Steps 50
HiDream-I1-Fast1024x10248+1-2Default3m 32.21sTest 13 - HiDream I1 Fast - steps and guidancedefaults Steps 16
HiDream-I1-Dev1024x102416+1-2Default12m 38.12sTest 14 - HiDream I1 Dev - steps and guidancedefaults Steps 28
stable-cascade1024x102420+2+Default2m 44.11sTest 4 - Stable Cascade - steps and guidance
SD3.5-large-turbo512x512
768x768		8+0-4Default
1m 18.23s		Test 15 - Stable Diffusion 3.5 Large Turbo - steps and guidancedefaults  CFG 0.0, Steps 4
Need to sign in and agree to the license
Cosmos-Predict2-14B1024x102416+2+Default23m 33.78sTest 17 - Cosmos Predict2 14B - steps and guidance

height and width have to be divisible by 16

Need to sign in and agree to the license

Kolors*1024x102416+3-8Default18m 39.49sTest 20 - Kolors - steps and guidancedefaults CFG=5, Steps=50

* Kolors executed on openvino, not ipex

SD.Next Supported Models

Sorted by model parameters

PublisherModelVersionSizeDiffusion ArchitectureModel ParamsText Encoder(s)TE ParamsAuto Encoder
HiDream-AIHiDreamI1 Fast/Dev/Full42.71 GB + 15.69MMDiT17.10BCLiP ViT-L + ViT+G + T5-XXL + LLama-3.1-8B0.12B + 0.69B + 2.95B + 4.54B16ch VAE
nVidiaCosmos-Predict2 T2I14B37.36GBMMDiT14.26BT5-XXL4.86WAN-VAE
Black Forest LabsFlux1 Dev/Schnell32.93GBMMDiT11.9BCLiP ViT-L + T5-XXL0.12B + 4.76B16ch VAE
Black Forest LabsFlux1 Kontext-Dev32.93GBMMDiT11.9BCLiP ViT-L + T5-XXL0.12B + 4.76B16ch VAE
StabilityAIStable Diffusion3.5 Large26.98GBMMDiT8.05BCLiP ViT-L + ViT+G + T5-XXL0.12B + 0.69B + 4.76B16ch VAE
FALAuraFlow0.331.90GBMMDiT6.8BUMT512.1BVAE
ThudmCogView430.39GBDiT6.37BGLM-49.40BVAE
StabilityAIStable CascadeMedium11.82GBMulti-stage UNet1.56B + 3.6BCLiP ViT-G0.69B42x VQE
NVLabsSana1.5 4.8B15.58GBMMDiT4.72BGemma22.61BDC-AE
OstrisFlex1 Alpha25.65GBMMDiT4.0BCLiP ViT-L + T5-XXL0.12B + 2.95B16ch VAE
VectorSpaceLabOmniGenv230.50GBTransformer3.97BQwen-VL-2.53.75BVAE
VectorSpaceLabOmniGenv115.47GBTransformer3.76BNone0VAE
KandinskyKandinsky327.72GBUnet3.05BT5-XXXL8.72BVQ
ThudmCogView3 Plus24.96GBDiT2.85BT5-XXL4.76BVAE
AlphaVLLMLumina220.75GBDiT2.61BGemma-22.61B16ch VAE
KwaiKolorsN/A17.40GBUNnet2.58BChatGLM6.24BVAE
StabilityAIStable DiffusionXL6.94GBUNet2.56BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
PlaygroundAIPlayground2.x13.35GBUNet2.56BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
StabilityAIStable Diffusion3.5 Medium15.89GBMMDiT2.25BCLiP ViT-L + ViT+G + T5-XXL0.12B + 0.69B + 4.76B16ch VAE
nVidiaCosmos-Predict2 T2I2B13.32GBMMDiT1.96BT5-XXL4.86WAN-VAE
Warp AIWuerstchenN/A12.16GBMulti-stage UNet1.0B + 1.05BCLiP ViT-L + ViT+G0.12B + 0.69B42x VQE
StabilityAIStable Diffusion3.0 Medium15.14GBMMDiT2.0BCLiP ViT-L + ViT+G + T5-XXL0.12B + 0.69B + 4.76B16ch VAE
StabilityAIStable CascadeLite4.97GBMulti-stage UNet0.7B + 1.0BCLiP ViT-G0.69B42x VQE
AlphaVLLMLuminaNext SFT8.67GBDiT1.7BGemma2.5BVAE
NVLabsSana1.5 1.6B9.49GBMMDiT1.60BGemma22.61BDC-AE
NVLabsSana1.0 1600M12.63GBMMDiT1.60BGemma22.61BDC-AE
DeepFloydIFL15.48GBMulti-stage UNet0.61B + 0.93BT5-XXL4.76BPixel
TencentHunyuanDiT1.214.09GBDiT1.5BBERT + T5-XL3.52B + 1.67BVAE
SegmindSSD-1BN/A8.72GBUNet1.33BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
KandinskyKandinsky2.25.15GBUnet1.25BCLiP ViT-G0.69BVQ
MeissonFlowMeissonicN/A3.64GBDiT1.18BCLiP ViT-H0.35BVQ
Thu-MLUniDiffuserv15.37GBU-ViT0.95BCLiP ViT-L + CLiP ViT-B0.12B + 0.16BVAE
StabilityAIStable Diffusion2.12.58GBUNet0.86BCLiP ViT-H0.34BVAE
StabilityAIStable Diffusion1.52.28GBUNet0.86BCLiP ViT-L0.12BVAE
PlaygroundAIPlayground14.95GBUNet0.86BCLiP ViT-L0.12BVAE
SalesforceBLIP-DiffusionN/A7.23GBUNet0.86BCLiP ViT-L + BLiP-20.12B + 0.49BVAE
DeepFloydIFM12.79GBMulti-stage UNet0.37B + 0.46BT5-XXL4.76BPixel
KoalaKoala700M6.58GBUNet0.78BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
SegmindVegaN/A6.43GBUNet0.75BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
PixArtAlphaXL 221.3GBDiT0.61BT5-XXL4.76BVAE
PixArtSigmaXL 221.3GBDiT0.61BT5-XXL4.76BVAE
Open-MUSEaMUSEd2563.41GBViT0.60BCLiP ViT-L0.12BVQ
NVLabsSana1.0 600M7.51GBMMDiT0.59BGemma22.61BDC-AE
SegmindTinyN/A1.03GBUNet0.32BCLiP ViT-L0.12BVAE
IDKiroSDXSN/A2.05GBUNet0.32BCLiP ViT-L0.12BVAE

from https://vladmandic.github.io/sdnext-docs/Models/

PublisherModelVersionSizeDiffusion ArchitectureModel ParamsText Encoder(s)TE ParamsAuto Encoder
StabilityAIStable Diffusion1.52.28GBUNet0.86BCLiP ViT-L0.12BVAE
StabilityAIStable Diffusion2.12.58GBUNet0.86BCLiP ViT-H0.34BVAE
StabilityAIStable DiffusionXL6.94GBUNet2.56BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
StabilityAIStable Diffusion3.0 Medium15.14GBMMDiT2.0BCLiP ViT-L + ViT+G + T5-XXL0.12B + 0.69B + 4.76B16ch VAE
StabilityAIStable Diffusion3.5 Medium15.89GBMMDiT2.25BCLiP ViT-L + ViT+G + T5-XXL0.12B + 0.69B + 4.76B16ch VAE
StabilityAIStable Diffusion3.5 Large26.98GBMMDiT8.05BCLiP ViT-L + ViT+G + T5-XXL0.12B + 0.69B + 4.76B16ch VAE
StabilityAIStable CascadeMedium11.82GBMulti-stage UNet1.56B + 3.6BCLiP ViT-G0.69B42x VQE
StabilityAIStable CascadeLite4.97GBMulti-stage UNet0.7B + 1.0BCLiP ViT-G0.69B42x VQE
Black Forest LabsFlux1 Dev/Schnell32.93GBMMDiT11.9BCLiP ViT-L + T5-XXL0.12B + 4.76B16ch VAE
Black Forest LabsFlux1 Kontext-Dev32.93GBMMDiT11.9BCLiP ViT-L + T5-XXL0.12B + 4.76B16ch VAE
OstrisFlex1 Alpha25.65GBMMDiT4.0BCLiP ViT-L + T5-XXL0.12B + 2.95B16ch VAE
NVLabsSana1.5 1.6B9.49GBMMDiT1.60BGemma22.61BDC-AE
NVLabsSana1.5 4.8B15.58GBMMDiT4.72BGemma22.61BDC-AE
NVLabsSana1.0 1600M12.63GBMMDiT1.60BGemma22.61BDC-AE
NVLabsSana1.0 600M7.51GBMMDiT0.59BGemma22.61BDC-AE
nVidiaCosmos-Predict2 T2I2B13.32GBMMDiT1.96BT5-XXL4.86WAN-VAE
nVidiaCosmos-Predict2 T2I14B37.36GBMMDiT14.26BT5-XXL4.86WAN-VAE
FALAuraFlow0.331.90GBMMDiT6.8BUMT512.1BVAE
AlphaVLLMLuminaNext SFT8.67GBDiT1.7BGemma2.5BVAE
AlphaVLLMLumina220.75GBDiT2.61BGemma-22.61B16ch VAE
PixArtAlphaXL 221.3GBDiT0.61BT5-XXL4.76BVAE
PixArtSigmaXL 221.3GBDiT0.61BT5-XXL4.76BVAE
SegmindSSD-1BN/A8.72GBUNet1.33BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
SegmindVegaN/A6.43GBUNet0.75BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
SegmindTinyN/A1.03GBUNet0.32BCLiP ViT-L0.12BVAE
KwaiKolorsN/A17.40GBUNnet2.58BChatGLM6.24BVAE
PlaygroundAIPlayground1.04.95GBUNet0.86BCLiP ViT-L0.12BVAE
PlaygroundAIPlayground2.x13.35GBUNet2.56BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
TencentHunyuanDiT1.214.09GBDiT1.5BBERT + T5-XL3.52B + 1.67BVAE
Warp AIWuerstchenN/A12.16GBMulti-stage UNet1.0B + 1.05BCLiP ViT-L + ViT+G0.12B + 0.69B42x VQE
KandinskyKandinsky2.25.15GBUnet1.25BCLiP ViT-G0.69BVQ
KandinskyKandinsky3.027.72GBUnet3.05BT5-XXXL8.72BVQ
ThudmCogView3 Plus24.96GBDiT2.85BT5-XXL4.76BVAE
ThudmCogView430.39GBDiT6.37BGLM-49.40BVAE
IDKiroSDXSN/A2.05GBUNet0.32BCLiP ViT-L0.12BVAE
Open-MUSEaMUSEd2563.41GBViT0.60BCLiP ViT-L0.12BVQ
KoalaKoala700M6.58GBUNet0.78BCLiP ViT-L + ViT+G0.12B + 0.69BVAE
Thu-MLUniDiffuserv15.37GBU-ViT0.95BCLiP ViT-L + CLiP ViT-B0.12B + 0.16BVAE
SalesforceBLIP-DiffusionN/A7.23GBUNet0.86BCLiP ViT-L + BLiP-20.12B + 0.49BVAE
DeepFloydIFM12.79GBMulti-stage UNet0.37B + 0.46BT5-XXL4.76BPixel
DeepFloydIFL15.48GBMulti-stage UNet0.61B + 0.93BT5-XXL4.76BPixel
MeissonFlowMeissonicN/A3.64GBDiT1.18BCLiP ViT-H0.35BVQ
VectorSpaceLabOmniGenv115.47GBTransformer3.76BPhi-30VAE
VectorSpaceLabOmniGenv230.50GBTransformer3.97BQwen-VL-2.53.75BVAE
HiDream-AIHiDreamI2 Fast/Dev/Full42.71 GB + 15.69MMDiT17.10BCLiP ViT-L + ViT+G + T5-XXL + LLama-3.1-8B0.12B + 0.69B + 2.95B + 4.54B16ch VAE

Size of the model on disk

Real disk size examples

root@server6:~# du -xsh /docker/sdnext/mnt-volume/models/Diffusers/* /docker/sdnext/mnt-volume/models/Stable-diffusion/*.safetensors |sort -n
2.0G    /docker/sdnext/mnt-volume/models/Stable-diffusion/dreamshaper_8.safetensors
2.0G    /docker/sdnext/mnt-volume/models/Stable-diffusion/epicrealism_naturalSin.safetensors
2.0G    /docker/sdnext/mnt-volume/models/Stable-diffusion/juggernaut_reborn.safetensors
2.0G    /docker/sdnext/mnt-volume/models/Stable-diffusion/majicmixRealistic_v7.safetensors
2.0G    /docker/sdnext/mnt-volume/models/Stable-diffusion/meinamix_v12Final.safetensors
3.3G    /docker/sdnext/mnt-volume/models/Diffusers/models--amused--amused-512
3.3G    /docker/sdnext/mnt-volume/models/Diffusers/models--stabilityai--stable-diffusion-x4-upscaler
3.5G    /docker/sdnext/mnt-volume/models/Diffusers/models--MeissonFlow--Meissonic
3.7G    /docker/sdnext/mnt-volume/models/Diffusers/models--segmind--SegMoE-SD-4x2-v0
4.0G    /docker/sdnext/mnt-volume/models/Stable-diffusion/counterfeitV30_25.safetensors
4.0G    /docker/sdnext/mnt-volume/models/Stable-diffusion/cyberrealistic_v80.safetensors
4.0K    /docker/sdnext/mnt-volume/models/Diffusers/models--PixArt-alpha--PixArt-Sigma-XL-2-2K-MS
4.0K    /docker/sdnext/mnt-volume/models/Diffusers/models--PixArt-alpha--PixArt-Sigma-XL-2-512-MS
4.8G    /docker/sdnext/mnt-volume/models/Stable-diffusion/stableDiffusion35_medium.safetensors
4.9G    /docker/sdnext/mnt-volume/models/Diffusers/models--stabilityai--stable-diffusion-2-1
4.9G    /docker/sdnext/mnt-volume/models/Stable-diffusion/kolors_v10.safetensors
5.2G    /docker/sdnext/mnt-volume/models/Diffusers/models--stablediffusionapi--anything-v5
5.2G    /docker/sdnext/mnt-volume/models/Diffusers/models--stablediffusionapi--disney-pixal-cartoon
5.2G    /docker/sdnext/mnt-volume/models/Diffusers/models--XCLiu--instaflow_0_9B_from_sd_1_5
5.4G    /docker/sdnext/mnt-volume/models/Diffusers/models--warp-ai--wuerstchen
6.2G    /docker/sdnext/mnt-volume/models/Diffusers/models--segmind--Segmind-Vega
6.3G    /docker/sdnext/mnt-volume/models/Diffusers/models--etri-vilab--koala-700m-llava-cap
6.3G    /docker/sdnext/mnt-volume/models/Diffusers/models--warp-diffusion--wuerstchen-prior
6.5G    /docker/sdnext/mnt-volume/models/Stable-diffusion/dreamshaperXL_v21TurboDPMSDE.safetensors
6.5G    /docker/sdnext/mnt-volume/models/Stable-diffusion/ponyDiffusionV6XL_v6TurboMerge.safetensors
6.5G    /docker/sdnext/mnt-volume/models/Stable-diffusion/ponyRealism_V23.safetensors
6.5G    /docker/sdnext/mnt-volume/models/Stable-diffusion/prefectPonyXL_v50.safetensors
6.5G    /docker/sdnext/mnt-volume/models/Stable-diffusion/realismByStableYogi_ponyV3VAE.safetensors
6.5G    /docker/sdnext/mnt-volume/models/Stable-diffusion/TempestV0.1-Artistic.safetensors
6.7G    /docker/sdnext/mnt-volume/models/Stable-diffusion/juggernautXL_ragnarokBy.safetensors
7.2G    /docker/sdnext/mnt-volume/models/Diffusers/models--Efficient-Large-Model--Sana_600M_512px_diffusers
8.1G    /docker/sdnext/mnt-volume/models/Diffusers/models--stabilityai--stable-diffusion-3.5-large-controlnet-depth
8.3G    /docker/sdnext/mnt-volume/models/Diffusers/models--Alpha-VLLM--Lumina-Next-SFT-diffusers
8.4G    /docker/sdnext/mnt-volume/models/Diffusers/models--segmind--SSD-1B
9.1G    /docker/sdnext/mnt-volume/models/Diffusers/models--Efficient-Large-Model--SANA1.5_1.6B_1024px_diffusers
9.1G    /docker/sdnext/mnt-volume/models/Diffusers/models--Efficient-Large-Model--Sana_1600M_1024px_BF16_diffusers
9.1G    /docker/sdnext/mnt-volume/models/Diffusers/models--Efficient-Large-Model--Sana_1600M_4Kpx_BF16_diffusers
9.1G    /docker/sdnext/mnt-volume/models/Diffusers/models--Efficient-Large-Model--Sana_Sprint_1.6B_1024px_diffusers
13G     /docker/sdnext/mnt-volume/models/Stable-diffusion/realvisxlV50_v50Bakedvae.safetensors
14G     /docker/sdnext/mnt-volume/models/Diffusers/models--stabilityai--stable-diffusion-xl-base-1.0
15G     /docker/sdnext/mnt-volume/models/Diffusers/models--Efficient-Large-Model--SANA1.5_4.8B_1024px_diffusers
15G     /docker/sdnext/mnt-volume/models/Diffusers/models--Shitao--OmniGen-v1
15G     /docker/sdnext/mnt-volume/models/Diffusers/models--stabilityai--stable-cascade
16G     /docker/sdnext/mnt-volume/models/Diffusers/models--tensorart--stable-diffusion-3.5-medium-turbo
16G     /docker/sdnext/mnt-volume/models/Stable-diffusion/stableDiffusion35_large.safetensors
17G     /docker/sdnext/mnt-volume/models/Diffusers/models--Kwai-Kolors--Kolors-diffusers
20G     /docker/sdnext/mnt-volume/models/Diffusers/models--Alpha-VLLM--Lumina-Image-2.0
21G     /docker/sdnext/mnt-volume/models/Diffusers/models--PixArt-alpha--PixArt-XL-2-512x512
24G     /docker/sdnext/mnt-volume/models/Diffusers/models--THUDM--CogView3-Plus-3B
25G     /docker/sdnext/mnt-volume/models/Diffusers/models--ostris--Flex.1-alpha
26G     /docker/sdnext/mnt-volume/models/Diffusers/models--stabilityai--stable-diffusion-3.5-large
26G     /docker/sdnext/mnt-volume/models/Diffusers/models--stabilityai--stable-diffusion-3.5-large-turbo
27G     /docker/sdnext/mnt-volume/models/Diffusers/models--kandinsky-community--kandinsky-3
28G     /docker/sdnext/mnt-volume/models/Diffusers/models--tensorart--stable-diffusion-3.5-large-TurboX
29G     /docker/sdnext/mnt-volume/models/Diffusers/models--THUDM--CogView4-6B
31G     /docker/sdnext/mnt-volume/models/Diffusers/models--stabilityai--stable-cascade-prior
32G     /docker/sdnext/mnt-volume/models/Diffusers/models--black-forest-labs--FLUX.1-dev
32G     /docker/sdnext/mnt-volume/models/Diffusers/models--black-forest-labs--FLUX.1-schnell
32G     /docker/sdnext/mnt-volume/models/Diffusers/models--shuttleai--shuttle-3.1-aesthetic
32G     /docker/sdnext/mnt-volume/models/Diffusers/models--shuttleai--shuttle-jaguar