stable-diffusion.cpp/examples/server/recipes/defaults.md

Built-in generation defaults

Reviewed against model-publisher cards and reference implementations on 2026-10-20.

These are starting values, not limits on supported resolutions or generation settings. Built-in step defaults above 20 are reduced to 20; the cited sources may recommend more steps. Users can still choose a higher step count. When several aspect ratios or ranges are published, the selected example or square bucket is noted below. A dash means no recipe override: the available primary sources did not establish a fixed value, or the size depends on the input image. Ordinary DiT recipes explicitly use Euler; PiD retains its LCM sampling path. Guidance-distilled models use cfg_scale: 1 to disable external CFG in sd.cpp. FLUX embedded guidance is stored separately as distilled_guidance. For community edit adapters and other optional extensions, these are the base-model settings unless the source specifies otherwise.

Recipe	CFG	Steps	Width × height	Source / notes
anima	4	20	1024 × 1024	Source The official ranges are 30–50 steps / CFG 4–5; CFG 4 selects the lower end. Edit uses the same base; tune the chosen editing LoRA separately.
anima-edit	4	20	1024 × 1024	Source The official ranges are 30–50 steps / CFG 4–5; CFG 4 selects the lower end. Edit uses the same base; tune the chosen editing LoRA separately.
animatediff-sd15	7.5	20	512 × 512	Source
boogu-image	4	20	1024 × 1024	Source
boogu-image-edit	5	20	1024 × 1024	Source
chroma	3	20	1024 × 1024	Source
chroma-radiance	3.5	20	1024 × 1024	Source
cosxl-edit	—	—	—	Source No numerical generation defaults in the author’s model card; previous test values were removed.
ernie-image	4	20	1024 × 1024	Source
ernie-image-turbo	1	8	1024 × 1024	Source
flex2	1	20	1024 × 1024	Source
flux1-canny	1	20	1024 × 1024	Source
flux1-depth	1	20	1024 × 1024	Source
flux1-dev	1	20	1024 × 1024	Source
flux1-fill	1	20	1232 × 1632	Source
flux1-kontext	1	20	—	Source Output size normally follows the reference image; no fixed size is imposed by this recipe.
flux1-pulid	1	20	1024 × 1024	Source FLUX.1-dev defaults; identity conditioning is provided by PuLID.
flux1-schnell	1	4	1024 × 1024	Source
flux2-dev	1	20	1024 × 1024	Source
flux2-klein-4b	1	4	1024 × 1024	Source
flux2-klein-9b	1	4	1024 × 1024	Source
flux2-klein-base-4b	4	20	1024 × 1024	Source
flux2-klein-base-9b	4	20	1024 × 1024	Source
flux3-klein	1	16	1024 × 1024	Source
flux3-klein-base	4	20	1024 × 1024	Source
hidream-o1-dev	1	20	2048 × 2048	Source
hunyuan-video-15-i2v	6	20	1280 × 720	Source 720p base I2V defaults; CFG/step-distilled and 480p checkpoints use different settings.
hunyuan-video-15-t2v	6	20	1280 × 720	Source 720p base T2V defaults; CFG-distilled / 480p checkpoints need their published settings.
ideogram4	7	20	1024 × 1024	Source Constant-guidance pipeline defaults; this recipe does not reproduce the CLI’s per-step guidance schedule.
instruct-pix2pix	—	10	—	Source Official quick-start example; output size follows the input image.
iris-3b	3	20	1024 × 1024	Source
krea2-raw	3.5	20	1024 × 1024	Source Base model defaults; edit requires a matching community LoRA with its own settings.
krea2-raw-edit	3.5	20	1024 × 1024	Source Base model defaults; edit requires a matching community LoRA with its own settings.
krea2-turbo	1	8	2048 × 2048	Source Official CLI example; edit requires a matching community LoRA.
krea2-turbo-edit	1	8	2048 × 2048	Source Official CLI example; edit requires a matching community LoRA.
lens	5	20	1440 × 1440	Source Official model card: RL model, default base_resolution 1440.
lens-turbo	1	4	1440 × 1440	Source Official model card: 4-step distilled model, default base_resolution 1440.
lingbot-video-dense	3	20	832 × 480	Source Official first-stage inference example; optional refiner is separate.
lingbot-video-moe	3	20	832 × 480	Source Official inference example.
llada-image	5	20	1024 × 1024	Source
llada-image-turbo	1	4	1024 × 1024	Source
longcat-image	4	20	1344 × 768	Source
longcat-image-edit	4.5	20	—	Source Reference-image-dependent output size; no fixed size is imposed.
ltx23-dev	3	20	768 × 512	Source Official single-stage parameters, before optional latent upscaling.
ltx23-distilled	1	8	768 × 512	Source Model card specifies 8 steps / CFG 1; size uses official pipeline’s first stage.
ltx25-dev	3	20	768 × 512	Source Official non-distilled pipeline parameters for model versions >= 2.4, before optional upscaling.
mage-flow	5	20	1024 × 1024	Source
mage-flow-edit	5	20	1024 × 1024	Source
mage-flow-turbo	1	4	1024 × 1024	Source
ming-image	1	12	2048 × 2048	Source
minimax-h3-fl2va	1	—	1344 × 768	Source Guidance-distilled; official 768p canvas. No fixed step count is prescribed here; the upstream API counts sigma grid points rather than sd.cpp denoising steps.
minimax-h3-ref2va	1	—	1344 × 768	Source Guidance-distilled; official 768p canvas. No fixed step count is prescribed here; the upstream API counts sigma grid points rather than sd.cpp denoising steps.
minit2i	2.5	20	512 × 512	Source B/16 example and native model size; L/16 uses CFG 6.
moss-image-base	3	20	1024 × 1024	Source
moss-image-sft	3	20	1024 × 1024	Source
nitrosd-realism	1	1	1024 × 1024	Source
ovis-image	5	20	1024 × 1024	Source
pid	1	4	2048 × 2048	Source Standalone PiD decoding at 2K; guidance is handled by the upstream generator.
pid-15	1	4	2048 × 2048	Source Standalone PiD decoding at 2K; guidance is handled by the upstream generator.
pid-15-flux2	1	4	2048 × 2048	Source Standalone PiD decoding at 2K; guidance is handled by the upstream generator.
pid-15-qwen-image	1	4	2048 × 2048	Source Standalone PiD decoding at 2K; guidance is handled by the upstream generator.
pid-15-sd3	1	4	2048 × 2048	Source Standalone PiD decoding at 2K; guidance is handled by the upstream generator.
pid-flux2	1	4	2048 × 2048	Source Standalone PiD decoding at 2K; guidance is handled by the upstream generator.
pid-qwen-image	1	4	2048 × 2048	Source Standalone PiD decoding at 2K; guidance is handled by the upstream generator.
pid-sd3	1	4	2048 × 2048	Source Standalone PiD decoding at 2K; guidance is handled by the upstream generator.
pixart-alpha-1024	4.5	20	1024 × 1024	Source
pixart-alpha-512	4.5	20	512 × 512	Source
pixart-sigma	4.5	20	1024 × 1024	Source
qwen-image	4	20	1328 × 1328	Source Official square resolution bucket; true_cfg_scale maps to cfg_scale.
qwen-image-21	1	20	2048 × 2048	Source Original checkpoint defaults; select the matching TAEQI2.1 to replace its VAE.
qwen-image-21-turbo	1	8	2048 × 2048	Model card, fixed sigmas. Standalone Turbo weights; the fixed 8-step schedule is applied on model load.
qwen-image-edit	4	20	—	Source Official editing example; output dimensions follow reference-image aspect ratio.
qwen-image-edit-2509	4	20	—	Source Official editing example; output dimensions follow reference-image aspect ratio.
qwen-image-edit-2511	4	20	—	Source Official editing example; output dimensions follow reference-image aspect ratio.
qwen-image-layered	4	20	640 × 640	Source Official recommended 640 resolution bucket; layered RGBA VAE is required.
sd-turbo	1	1	512 × 512	Source
sd1-checkpoint	7.5	20	512 × 512	Source
sd1-controlnet	7.5	20	512 × 512	Source
sd1-inpainting	7.5	20	512 × 512	Source
sd1-ip-adapter	7.5	20	512 × 512	Source SD 1.x base pipeline defaults; adapter strength is configured separately.
sd2-512-checkpoint	7.5	20	512 × 512	Source
sd2-768-checkpoint	7.5	20	768 × 768	Source
sd2-inpainting	7.5	20	512 × 512	Source
sd3-medium	7	20	1024 × 1024	Source
sd35-large	3.5	20	1024 × 1024	Source
sd35-medium	4.5	20	1024 × 1024	Source
sdxl-checkpoint	5	20	1024 × 1024	Source
sdxl-controlnet	5	20	1024 × 1024	Source SDXL pipeline defaults; compatible ControlNet weights are required.
sdxl-inpainting	8	20	1024 × 1024	Source
sdxl-ip-adapter	5	20	1024 × 1024	Source SDXL base pipeline defaults; adapter strength is configured separately.
sdxl-photomaker	—	—	1024 × 1024	Source Use settings for the selected compatible SDXL checkpoint; no unique CFG/steps are prescribed by the model card.
sdxl-turbo	1	1	512 × 512	Source
sdxs	1	1	512 × 512	Source
sefi-image	4	20	—	Source Base/RL family defaults; Turbo checkpoints require 4 steps / CFG 1.
sensenova-u15	4	20	2048 × 2048	Source
wan21-flf2v	5	20	1280 × 720	Source
wan21-i2v	5	20	832 × 480	Source 480p base defaults; also accepts compatible SkyReels V2 and single Wan 2.2 I2V experts.
wan21-t2v	5	20	1280 × 720	Source 14B base defaults; single Wan 2.2 experts and acceleration LoRAs need matching settings.
wan21-vace	5	20	832 × 480	Source
wan22-i2v	3.5	20	1280 × 720	Source
wan22-s2v	4.5	20	—	Source The official S2V pipeline derives output dimensions from the input.
wan22-t2v	3	20	1280 × 720	Source
wan22-ti2v-5b	5	20	1280 × 704	Source
z-image	4	20	720 × 1280	Source
z-image-l2p	—	—	—	Source No numerical inference recommendations in the author’s model card; previous test values were removed.
z-image-turbo	1	9	1024 × 1024	Source

TAE alternatives use the model-compatible weights listed by TAESD and TAEHV. TAE slots are omitted for pixel-space models, PiD, layered RGBA decoding and Mage-VAE paths that the current runtime does not map to a compatible TAE.