NightCafe AI Models Guide

NightCafe AI Models Guide: Choosing the Right Model

Not sure which model to use? Start here

NightCafe offers a lot of models, and that can feel like a lot to take in when you're starting out. The good news is that you don't need to work your way through the list. A small group of models covers almost everything, and these are the ones we'd recommend reaching for first:

If you want…Use thisApprox. credits
One model for almost everything — best everyday valueGPT Image 2.5 Flare on Quality Low (PRO)~1
Newest OpenAI generation, when you will spend more than 1 creditGPT Image 2.6 Flare or Sunburst on Quality Low (PRO). Sunburst for detailed scenes; Flare to iterate faster. Same price.2.5
Flagship quality for very few creditsMuse Image (PRO) — unlimited on AI Creator+~2
Flux 3 quality for fewer credits (PRO)Flux 3 Klein — 1K to 4K plus reference-image editingfrom 0.75
Realistic Seedream, 1K (PRO)Seedream 5.5 — 2 credits at 1K; 0.75 per reference; transparent PNG. Flash stays a flat 2 at 2K2
More detail and polish from that same familyGPT Image 2.5 Quality Medium / High, or GPT Image 2 Medium / High~1.5+ / ~5.5 / ~22
The strongest image editing available in StudioGPT Image 2.5, GPT Image 2, Muse Image~1+ / ~1+ / ~2
Good results without a PRO subscriptionZ-Image Turbo, or Flux 2 Klein 9B Fast if you also want to edit images~0.5 / ~0.75
Completely free (no credits)Stable Diffusion 1.5 or DreamShaper v80
Short video on any PRO plan for few creditsP-Video-2 Fast (PRO)~8 for 5s at 720p

If you only remember one thing: GPT Image 2.5 Flare on Quality Low is the default recommendation for most people and most prompts. It costs about one credit, and is unlimited on AI Creator and above at Low (up to 2 credits — so Low at Medium Res or High Res). GPT Image 2.6 launched on 9 November 2026 and now sits above 2.5 in the image picker. Its default Low setting is 2.5 credits, and it is not unlimited. Use 2.6 when you want that newer model. See GPT Image 2.6 on NightCafe and GPT Image 2.5 on NightCafe.

The models above are the safe choices. Every other model has its own look, and the rest of this guide covers which one to pick when you want something specific.

If you would rather see models side by side, NightCafe has a free AI Model Comparison at creator.nightcafe.studio/evals — including an Image Editing tab for prompt-to-edit. Looking at it does not spend credits. You can also open it from Studio's model picker, Prompt to Edit, your profile menu, or the site footer.

NightCafe 101: Create Your First Image aired on 2 November 2026. The page has the replay and details.

How to read this guide

  • Newer models often sit higher in the picker. That is not the same as the best credit value. The start of this guide names the everyday pick separately.
  • Approx. credits is the base cost for one standard image. The final cost changes with resolution, number of images, reference images, and any PRO unlimited allowances.
  • PRO means the model needs a NightCafe PRO subscription. Everything marked otherwise can be used on a free account with credits.
  • Fast credits: a few premium models (GPT Image 2 High, Nano Banana Pro) can only be run with purchased "fast" credits, not earned credits in relax mode.
  • Editing terms: "prompt-to-edit" = change an existing image with a text instruction. "Multi-image fusion" = combine several reference images into one result. "Consistent character" = keep the same character or subject across images (comics, storyboards).

Recommendations below are based on the Artificial Analysis image arenas (blind human preference voting) cross-referenced against NightCafe's own credit pricing. Most text-to-image ranks in this edition were checked on 9 November 2026. Seedream ranks were checked on 13 November 2026. The Seedance 2.5 video rank was checked on 16 November 2026.

Best value for money

"Value" means quality per credit, not simply the cheapest option. Because the strongest value picks are PRO models, the lists are split.

Best value — PRO models

ModelApprox. creditsWhy it's good value
GPT Image 2.5 Flare / Sunburst (Low)~1The best everyday value on NightCafe. Quality Low by default, Prompt to Edit, multi-image fusion and consistent characters. Flare is the faster everyday pick; Sunburst is the higher-precision twin at the same price. Unlimited on AI Creator+ when the job is 2 credits or less. GPT Image 2.6 is newer, and its default is 2.5 credits.
GPT Image 2.6 Flare / Sunburst (Low)2.5Newest OpenAI generation (9 November 2026). Same Quality ladder, prompt-to-edit, fusion and consistent characters as 2.5, at a higher credit cost. Independent text-to-image arenas rank Sunburst (max) first and Flare (max) second (9 November 2026). Those scores are Max quality (21.25 credits at about 1K). Not unlimited. See GPT Image 2.6 on NightCafe.
Flux 3 Kleinfrom 0.75The affordable Flux 3 (PRO). 1K from 0.75 credits, 2K from 1.75, 4K from 10. Prompt-to-edit and up to 4 reference images (extra cost capped after two). Independent text-to-image arenas list FLUX 3 Image Klein around 24th (Elo about 1043, as of 9 November 2026). It is a value model, not a top rank, and it is the cheapest way to get Flux 3 generation and 4K on NightCafe. See Flux 3 Klein on NightCafe.
Muse Image~2Meta's flagship. Strong text-to-image and editing for two credits, and a single image is unlimited on AI Creator and above. The best quality-per-credit above the sub-1-credit tier if you want a 2K default.
Seedream 5.52Newer realistic Seedream (13 November 2026). 2 credits at 1K, 3.5 at 2K, plus 0.75 per reference (up to 10). Transparent PNG. Text-to-image arenas: about 10th, Elo about 1134 (13 November 2026). See Seedream 5.5 on NightCafe.
Seedream 5.0 Flash2Flat 2 credits at 1K or 2K, references included. Arenas have not listed it. See Seedream 5.0 Flash on NightCafe.
GPT Image 2 Low~1The previous OpenAI generation, still excellent. Independent text-to-image arenas rank GPT Image 2.6 (max) first, with GPT Image 2.5 (max) just behind (9 November 2026). Low is the cheap setting of the previous-generation family. Unlimited on AI Creator+.
Qwen Image 2512~1Strong realism and excellent text for a single credit (text-to-image only).
GPT Image 1.5 Low~1The previous OpenAI generation. It shows Gone Soon and leaves on December 1, 2026. Prefer GPT Image 2.5 Low or GPT Image 2 Low.
Nano Banana 2.1~3.5 at 1KUpdated Google model (October 2026): 1K–4K, up to 14 reference images, prompt-to-edit. Unlimited 1K on AI Artist. See Nano Banana 2.1 on NightCafe.
Nano Banana 2 Lite~3.5A simpler 1K-only Google editor if you do not need 2.1's extra resolution or 14-image fusion.
Seedream 4.0 / 4.5~3.5 / ~4.5Still the Seedream pick for native 4K. For 1K realism use Seedream 5.5; Flash keeps 2K at 2 credits.

Best value — no PRO subscription needed

ModelApprox. creditsWhy it's good value
Z-Image Turbo~0.5The best realism per credit without a subscription. Text-to-image only.
Flux 2 Klein 9B Fast~0.75The best all-rounder on a free account: sub-second generation plus prompt-to-edit, multi-image fusion and consistent characters.
Flux Schnell~0.5Very fast and very cheap, at older Flux quality.
Flux 2 Klein 4B Fast~0.5A smaller Klein model — fine for drafts and iteration.
HiDream I1 Fast~1Artistic, high-clarity looks at low cost.
Qwen Image Edit Plus~2The best photo editing available without PRO; also combines reference images.
Flux Kontext Dev~2.5Natural-language photo editing, no PRO required.
Stable Diffusion 1.5, DreamShaper v8FreeFree base generations, at 2022-era quality.

Best model for each task

A fast lookup. Picks are deliberately short — the first model listed is the one to suggest unless the person has a reason to avoid it.

Task / use-caseBest picksWhy
Anything and everythingGPT Image 2.5 Flare (Low), then raise Quality or try SunburstBest everyday model at ~1 credit, and unlimited on AI Creator and above at Low. GPT Image 2.6 is newer and costs 2.5 credits at Low. Independent text-to-image arenas rank 2.6 Max first; those scores are the expensive Max setting.
Best value (PRO)GPT Image 2.5 Low, Flux 3 Klein, Muse Image, Seedream 5.5, GPT Image 2 LowFlagship-class quality for 1–2 credits (Klein from 0.75), with Creator+ unlimited on GPT 2.5 Low / Muse singles. Seedream 5.5 is 2 credits at 1K.
Best value (no PRO)Z-Image Turbo, Flux 2 Klein 9B FastStrong results for well under a credit.
FreeStable Diffusion 1.5, DreamShaper v8Free base generations, no PRO required.
Editing an existing imageGPT Image 2.6, GPT Image 2.5, GPT Image 2, Muse Image, Nano Banana 2.1The strongest editors currently in the Studio picker. Independent arenas rank Nano Banana 2.1 near the top for text-to-image (about 6th, Elo about 1176, as of 9 November 2026) at ~3.5 credits for 1K. Flux 3 Klein is the cheap PRO Flux editor (from 1.5 credits for an edit). Seedream 5.5 edits are 2.75 credits (1K source) or 4.25 (2K). Seedream 5.0 Flash stays a flat 2. Flux 2 Klein 9B Fast is the non-PRO option. Compare editors on the Image Editing tab.
PhotorealismGPT Image 2.5, GPT Image 2, Grok Image 2.0 MediumLifelike people, skin, lighting and natural detail.
Text & typographyGPT Image 2.5, Flux 3, GPT Image 2, Nano Banana 2.1 / Pro, Qwen Image 3.0 ProAccurate spelling and layout for posters, signs and captions.
Complex prompts / prompt adherenceGPT Image 2.5 Sunburst (higher Quality), GPT Image 2 High, Nano Banana 2.1 / ProFollows long, detailed instructions faithfully.
Artistic & stylised workGPT Image 2.5, GPT Image 2 Medium, Muse Image, Seedream 5.0 Flash, Seedream 4.xPainterly, stylised and concept-art looks.
Logos, graphic design & vectorsRecraft v3, Ideogram V3 Quality, Flux 2 FlexClean design output; Recraft v3 can generate scalable vectors.
Combining reference imagesGPT Image 2.5, Flux 3, Flux 3 Klein, Seedream 5.5, Seedream 5.0 Flash, GPT Image 2, Muse Image, Nano Banana 2.1Merge several inputs (e.g. product + scene) into one image. Nano Banana 2.1 takes up to 14 references; Flux 3, Seedream 5.5, and Seedream 5.0 Flash take up to 10. Flash includes them in the 2-credit price; 5.5 charges 0.75 each. Flux 3 Klein takes up to 4 (extra credits, capped after two).
Consistent characters (comics, storyboards)GPT Image 2.5, GPT Image 2, Muse Image, Seedream 5.5, Nano Banana 2.1Keep the same character across multiple images.
Anime & illustrationAnimagine XL, Blue Pencil XL, AAM XL Anime MixDedicated anime checkpoints; Seedream for modern stylised looks.
Highest resolution (4K)Flux 3 Klein, Flux 3, Nano Banana 2.1, Seedream 4.0 / 4.5Native 4K and other large, print-ready output. Flux 3 Klein is the cheaper 4K Flux option (from 10 credits); full Flux 3 Ultra is the premium 4K setting.
VideoP-Video-2 Fast, MiniMax H3 Max / H3 Max Turbo, Seedance 2.5 Flash, Seedance 2.5, Kling 4.0 Flash, Kling 4.0, PixVerse V6 (Studio default), Google Veo 3.1Text-to-video and image-to-video with native audio. P-Video-2 Fast is the cheapest current any-PRO option (~8 credits for 5s at 720p). Gemini Omni Flash 1.2 (9 November 2026) is the current Google Omni Flash: up to 15 seconds, 360p to 4K, 27 credits for 5 seconds at 720p. H3 Max ranks at the top of independent image-to-video-with-audio arenas; H3 Max Turbo is the quality-per-credit pick in that family. Seedance 2.5 Flash (16 November 2026) is the cheaper 480p/720p Seedance (58 credits for 5s at 720p). Seedance 2.5 keeps 1080p. New videos in Studio start on PixVerse V6. Kling 4.0 Flash is the cheaper 720p Kling; Kling 4.0 adds 1080p and longer clips. Credit figures are in the video-model list below.
Your own face, object or styleFine-tuned LoRA modelsTrain a model on your own images — see the LoRA section.

These are the models at the top of the model picker, described in more detail. Anything not listed in this section is still available, but there is usually a better or cheaper option here.

OpenAI — GPT Image 2.6, 2.5, 2 and 1.5

GPT Image 2.6 is the newest OpenAI family on NightCafe (9 November 2026). It ships as two PRO models that share a Quality setting (Low through Max) and Medium Res / High Res: Sunburst (detailed scenes and tighter control) and Flare (faster iteration). They cost the same. Both support prompt-to-edit, multi-image fusion (up to 16 start images) and consistent characters. Low at Medium Res is 2.5 credits. It is not an unlimited model. GPT Image 2.6 on NightCafe.

GPT Image 2.5 is the previous generation and still the one to start with for everyday work. Flare on Quality Low is about 1 credit and is unlimited on AI Creator and above when the job is 2 credits or less. Sunburst uses the same credit table when you want tighter control. GPT Image 2.5 on NightCafe.

GPT Image 2 is still available and still excellent. Independent Artificial Analysis text-to-image arenas (9 November 2026) rank GPT Image 2.6 Sunburst (max) first (Elo about 1259) and Flare (max) second (Elo about 1228), with GPT Image 2.5 Sunburst (max) third. Those arena scores are for the Max quality setting — 21.25 credits at about 1K on 2.6, the expensive end of the Quality ladder. For everyday use, GPT Image 2.5 Flare on Quality Low is still the value pick at about 1 credit. The Low / Medium / High tiers of GPT Image 2 are the same previous-generation model at different quality settings.

  • GPT Image 2 Low (~1 credit, PRO) — still a strong one-credit option; unlimited on AI Creator+.
  • GPT Image 2 Medium (~5.5 credits, PRO) — step up when Low isn't detailed enough.
  • GPT Image 2 High (~22 credits, PRO, fast credits only) — the highest-fidelity GPT Image 2 setting.
  • GPT Image 1.5 (Low / Medium / High) (~1 / 3.5 / 10 credits, PRO) — shows Gone Soon and will be removed on December 1, 2026. 1.5 Low stays unlimited on AI Artist+ until then. See Deprecation of GPT Image 1.5.
  • GPT Image 1 (Low / Medium / High) — superseded by 1.5, 2 and 2.5. Retired on October 23, 2026 and no longer available for new creations. See Deprecation of GPT Image 1.

Meta — Muse Image

Muse Image (~2 credits, PRO) is Meta's flagship image model and one of the best value models on NightCafe. A single image is unlimited on AI Creator and above (4-image batches still cost credits). It is near the top of independent text-to-image arenas, with prompt-faithful generation, precise edits, multi-reference composition (up to 10 start images), consistent characters and strong text rendering at 2K. Recommend it whenever someone wants flagship-class quality at two credits. See Muse Image on NightCafe.

Microsoft — MAI Image

Microsoft's MAI Image models score very highly on independent arenas (MAI Image 2.6 is typically just behind GPT Image 2 High for text-to-image). They are not currently in the NightCafe Studio model picker, so do not recommend them as something you can generate with today. Use GPT Image 2.5, GPT Image 2 or Muse Image instead.

Google — Nano Banana (Gemini)

Google's Gemini image models (nicknamed "Nano Banana") have first-class start-image support and are excellent for editing and combining photos, with very strong text rendering.

  • Nano Banana 2.1 (~3.5 credits at 1K, PRO) — Google's updated Nano Banana 2, launched October 2026. 1K–4K output, up to 14 reference images included in the output price, prompt-to-edit, fusion and consistent characters. Unlimited 1K generations on AI Artist. Independent Artificial Analysis text-to-image boards list Nano Banana 2.1 about 6th (Elo about 1176, as of 9 November 2026). On NightCafe this is the Nano Banana to start with. See Nano Banana 2.1 on NightCafe.
  • Nano Banana 2 Lite (~3.5 credits, PRO) — a faster, 1K-only variant for everyday text-to-image and editing.
  • Nano Banana 2 (~7 credits at 2K, PRO) — AKA Gemini Flash 3.1. Still available, with 4K support, fusion and consistent characters. Prefer Nano Banana 2.1 for new work — it is cheaper at 1K and 4K and takes more reference images.
  • Nano Banana Pro (~11 credits, PRO, fast credits only) — the most expensive Nano Banana, with very high prompt-adherence and typography scores.
  • Nano Banana (Gemini Flash 2.5) — retired on October 2, 2026 and no longer available for new creations. See Deprecation of Nano Banana. Use Nano Banana 2.1 instead.

xAI — Grok Imagine

  • Grok Image 2.0 Low (~4 credits, PRO) — xAI's latest generation on the faster setting, with 1K/2K output and prompt-to-edit from up to three reference images. Early editing-arena results for Grok 2.0 are very strong.
  • Grok Image 2.0 Medium (~6 credits, PRO) — the default flagship quality for Grok 2.0.
  • Grok Imagine Image Quality (~7 credits, PRO) — previous generation. Retired on November 2, 2026. It is no longer available for new creations. See Deprecation of Grok Imagine Image Quality. Switch to Grok Image 2.0 Medium.
  • Grok Imagine Image (~2 credits, PRO) — the previous generation, still available.

ByteDance — Seedream

Seedream models are fast and high-resolution, and support editing and fusion. Seedream 5.5 is the newer realistic Seedream (2 credits at 1K). Seedream 5.0 Flash stays the flat 2-credit option, including at 2K and with references included. Seedream 4.0 / 4.5 remain the go-to for native 4K. Family overview: Seedream on NightCafe.

  • Seedream 5.5 (2 credits at 1K, 3.5 at 2K, PRO) — launched 13 November 2026. Prompt-to-edit, up to 10 references at 0.75 credits each, transparent PNG, up to 4 images. Edits are 2.75 (1K) or 4.25 (2K) and stay opaque. Text-to-image arenas: about 10th, Elo about 1134 (13 November 2026). Seedream 5.5 on NightCafe.
  • Seedream 5.0 Flash (2 credits, PRO) — 21 October 2026. Flat 2 credits at 1K or 2K, references included. Arenas have not listed it. Seedream 5.0 Flash on NightCafe.
  • Seedream 5.0 Pro (~5 credits, PRO) — older 2K option. Arenas place it about 19th (Elo about 1103, 13 November 2026), behind 5.5.
  • Seedream 5.0 Lite (Preview) (~3.5 credits, PRO) — up to 3K.
  • Seedream 4.5 / 4.0 (~4.5 / ~3.5, PRO) — native 4K Seedreams.

Black Forest Labs — Flux

The Flux family spans near-free distilled models to high-end production tools, and includes dedicated editing models (Kontext). Flux 3 is the newest PRO flagship (native 4K). Flux 3 Klein is the fast, affordable Flux 3 (also PRO). The Flux 2 Klein models are still the reason to come here on a free account — they remain the best non-PRO options on NightCafe.

  • Flux 3 Klein (from 0.75 credits at 1K, PRO) — the fast, affordable Flux 3, launched 20 October 2026. Text-to-image up to 4K, prompt-to-edit, up to 4 reference images, consistent characters, and a Seed control. 1K / 2K / 4K cost 0.75 / 1.75 / 10 credits with no references (references add a capped surcharge). Independent text-to-image arenas list it around 24th (Elo about 1043, as of 9 November 2026) — a value model, not a top rank. Flux 3 Klein on NightCafe.
  • Flux 3 (~4 / 4.75 / 10 / 60.75 credits at Low / Medium / High / Ultra, PRO) — Black Forest Labs' flagship image model (October 2026). Native 4K, prompt-to-edit, up to 10 reference images included in the output price, consistent characters, and accurate text you can change in place. Default is Medium (~1K, 4.75 credits). Independent Artificial Analysis text-to-image arenas list FLUX 3 Image around 15th (Elo about 1117, as of 9 November 2026) — a strong high-resolution option, not the everyday 1-credit pick. Prefer Flux 3 Klein when you want Flux 3 at a lower credit cost. Flux 3 on NightCafe.
  • Flux 2 Klein 9B Fast (~0.75 credits, no PRO) — sub-second generation plus editing, fusion and consistent characters. The recommended free-account model. Prefer Flux 3 Klein if you have PRO.
  • Flux 2 Klein 9B (~1 credit, PRO) and Klein 4B / 4B Fast (~1 / 0.5 credits, no PRO) — other points on the same speed/quality curve.
  • Flux 2 Max (~7 credits, PRO), Flux 2 Pro (~3), Flux 2 Flex (~6, tuned for typography), Flux 2 Dev (~1.5) — the previous Flux 2 line, all with editing, fusion and consistent characters. Flux 3 is the newer high-end pick when you want native 4K.
  • Flux Kontext (Dev / Pro / Max) (~2.5 / 3.5 / 7 credits) — specialised for editing images with natural-language prompts. Kontext Dev needs no PRO.
  • Flux Schnell (~0.5 credits, no PRO), Flux (~1, PRO), Flux Krea (~1.5, PRO), Flux PRO 1.1 (~4) and 1.1 Ultra (~7, 2K) — the older Flux 1 generation.
  • Juggernaut Flux (Base / Pro) — realism-focused Flux fine-tunes by RunDiffusion.

Alibaba — Qwen Image & Z-Image

  • Z-Image Turbo (~0.5 credits, no PRO) — excellent realism for the price and the best free-account text-to-image model.
  • Qwen Image 2512 (~1 credit, PRO) — natural humans, sharp textures, accurate text for one credit.
  • Qwen Image 3.0 / 3.0 Pro (~3.5 / 7.5 credits, PRO) — the latest Qwen: native 2K, class-leading typography, multi-image editing. 3.0 Pro is top-15 worldwide for both generation and editing.
  • Qwen Image 2.0 / 2.0 Pro (~3.5 / 7.5 credits, PRO) — the previous unified generation-and-editing models.
  • Qwen Image Edit / Edit Plus / Edit 2511 (~1 / 2 / 2 credits) — dedicated editing models. Edit and Edit Plus need no PRO.
  • Qwen Image / Qwen Image SD (~1 / 0.75 credits) — typography-strong base models; SD is cheaper and lower-res.

Other current models

  • HiDream I1 Dev / Full / Fast (~1.5 / 2 / 1 credits) — artistic, high-clarity generation at different speed and cost points. I1 Fast needs no PRO.
  • Ideogram V3 / V3 Quality / V3 Turbo (~4 / 6 / 2 credits, PRO) — strong prompt adherence and text-in-image. Earlier Ideogram 2.x models are also still available.
  • Recraft v3 (~5 credits, PRO) — excellent spelling and typography, and the only model that can output vector graphics.
  • Stable Core (~4 credits, PRO) — beautiful high-resolution images without complex prompts.

Video models

Video models generate short clips from a text prompt and/or a start image, and many include synchronised audio. All video models are PRO, most require a higher PRO plan level, and credit costs are per clip and much higher than image models. Studio starts new videos on PixVerse V6 by default; you can switch to any other current video model in the picker. Provider chips (including MiniMax and Pruna AI) help you jump to a family.

  • P-Video-2 Fast / P-Video-2 / P-Video-2 Pro Fast / P-Video-2 Pro — Pruna AI's NightCafe video family (28 September 2026). Fast and standard: any PRO plan, regular credits, 720p or 1080p, up to 20 seconds, optional native audio and start/end frames. Pro / Pro Fast: Enthusiast+, Fast Credits, 480p or 768p, 5–15 seconds, audio always on. P-Video-2 Fast is the cheapest current everyday video option (~8 credits for 5s at 720p). P-Video-2 and P-Video-2 Pro on NightCafe.
  • Google Veo 3.1 / 3.1 Fast / 3.1 Lite — lifelike motion with context-aware synchronised audio and first/last-frame references. Veo 3.1 is the most popular pick.
  • Seedance 2.5 Flash — cheaper 480p/720p Seedance (16 November 2026). No 1080p. 5 seconds is 26 credits at 480p or 58 at 720p. Seedance 2.5 Flash.
  • Seedance 2.5 — the 1080p Seedance. About 4th on Artificial Analysis text-to-video on 16 November 2026 (Elo about 1156). Seedance 2.5.
  • Seedance 2.0 / 2.0 Fast / 2.0 Mini — earlier Seedance models for shorter clips. Seedance 1.5 Pro was removed on November 11, 2026. See Deprecation of Seedance 1.5 Pro.
  • Kling 4.0 Flash — the cheaper 720p Kling, launched 5 November 2026. Text, an optional start image, or up to 9 reference images. 720p only, 3–20 seconds, audio always on. Enthusiast+ and fast credits. 7.5 credits a second (38 credits for 5 seconds, 150 for 20 seconds). It does not do 1080p, clips longer than 20 seconds, or an end frame. Artificial Analysis has not ranked Kling 4.0 Flash (checked 9 November 2026). Kling 4.0 Flash on NightCafe.
  • Kling 4.0 — the longer Kling, launched 29 October 2026. Same inputs as Flash, plus 1080p and clips up to 30 seconds. A 5-second clip is 59 credits at 720p and 89 at 1080p. Choose this when you need 1080p or a clip longer than 20 seconds. Kling 4.0 on NightCafe. Kling 3.0 Turbo (up to 15 seconds) and Kling 2.5 Turbo Pro / Standard are the older Kling options. 2.5 Turbo Pro can lock an end frame; Kling 4.0 and Flash cannot.
  • Wan 3.0 / 3.0 Prime — Alibaba's latest video models.
  • Grok Imagine Video — xAI's multimodal video with native audio from text or a start image.
  • PixVerse V6 (Studio default) / V5 — multi-shot storytelling with synchronised audio and cinematic controls.
  • Happy Horse 1.0 — best-in-class motion quality with synchronised audio.
  • Gemini Omni Flash 1.2 — Google's current Omni Flash, launched 9 November 2026. Native audio, 3–15 seconds, 360p through 4K, start and end frames, or up to 10 reference images. 16:9 or 9:16. Enthusiast+ and fast credits. A default 5-second 720p clip is 27 credits (5.25 credits a second, rounded up). An end frame or reference images cap the clip at 10 seconds, and reference images cannot be combined with a start or end frame. Gemini Omni Flash 1.2 on NightCafe. Gemini Omni Flash 1.1 shows Gone Soon and leaves December 3, 2026. See Deprecation of Gemini Omni Flash 1.1.
  • MiniMax H3 Max / H3 Max Turbo / H3 Fast — MiniMax's newest NightCafe family (September 2026). All produce 5–15 second clips with native audio. Independent Artificial Analysis video arenas (with audio, as of September 2026) rank H3 Max first for image-to-video and near the top for text-to-video. H3 Max Turbo is the value pick in that family: any PRO plan, about half the credits of H3 Max, 480p or 720p. H3 Max adds up to nine reference images (or a start/end frame) and needs Enthusiast+ plus Fast Credits. H3 Fast is a 480p preview/batch option (Enthusiast+, Fast Credits). The earlier MiniMax H3 2K model is still available but costs more per second. MiniMax H3 Max, Turbo and Fast on NightCafe.
  • Seedance 1.0 Pro / Pro Fast — earlier Seedance models, still available.

LoRAs & fine-tuned (custom) models

In addition to the official models above, you can train your own custom model on NightCafe. These are LoRAs ("Low-Rank Adaptation") — a fast way to fine-tune Stable Diffusion on a small set of images so it can recreate a specific face, object/animal, or style.

How to train a LoRA

  1. Open My Models (in the main menu on desktop, or under your profile picture on mobile) and click Fine-tune a new model.
  2. Choose a model type (face, object/animal, or style) and give it a name.
  3. Choose or create a dataset — upload at least ~20 images. More varied images generally give better results.
  4. Agree to the terms and start training. Training typically takes about 10–30 minutes, and you'll be notified when it's done.

Fine-tuning is a PRO feature, but free users get 1 free face-model tune plus 10 free generations with it. Your trained models are private to your account.

How to use a LoRA in a prompt

To use a fine-tuned model, add its token to your prompt in the format <type:name:optional weight> — for example <lora:My Face:0.8>. You can write prompts like "A photo of <lora:My Face:0.8> riding an elephant" or "A unicorn in the style of <lora:Dark Fantasy:0.5>". The weight is optional (usually between 0 and 1) and defaults to 0.8 if omitted.

Learn more in Introducing Fine-Tuning on NightCafe.

Older model classes (still available)

NightCafe keeps many earlier open-source checkpoints available. They are cheap and can be great for specific styles, but they have much lower prompt adherence and text ability than anything in the recommended list above. Rather than list every checkpoint, they're grouped into classes below.

SDXL models (Stable Diffusion XL)

SDXL launched in 2023 as a major step up from SD 1.5, producing sharper 1024px images. NightCafe hosts many community SDXL checkpoints, each tuned for a look. All are low-cost (~0.5–1 credit), and "Lightning" variants are the fastest and cheapest. Examples by strength:

  • Photorealism: RealVisXL (v3–v5), Crystal Clear XL, Cherry Picker XL, Fluently XL, Juggernaut XL / v9 / XI.
  • Art & concept: DreamShaper XL, Starlight XL, Mysterious XL.
  • Anime & stylised: Animagine XL, Blue Pencil XL, AAM XL Anime Mix, Virtual Utopia XL.
  • Cartoon: Real Cartoon XL.
  • Base & utility: SDXL 1.0, SDXL DPO.

SD 1.5 models (Stable Diffusion 1.5)

SD 1.5 is the first-generation Stable Diffusion model from 2022. It's fast and inexpensive (some are free), but lower resolution and far less accurate than modern models. Stable Diffusion v1.4, the original 2022 model, is in the main model list too. It is free, low resolution, and not a model to reach for when a newer one will do. NightCafe hosts many SD 1.5 community checkpoints, including:

  • Realism: Realistic Vision, AbsoluteReality, Juggernaut Reborn.
  • Art & "does-it-all": DreamShaper v8 (free), NeverEnding Dream.
  • Anime, comics & cartoon: Blue Pencil, Arthemy Comics, 3D Animation Diffusion, RealCartoon Pixar, Ghost Mix.
  • Niche: RPG (character portraits), Rabbit (cute art), WildlifeX (animals), Nightmare Shaper (dark art).

Upscalers

Open a finished creation and choose Upscale. The default regular upscaler is now P Image Upscale (2×–16×, with Enhance detail / Enhance realism). Real-ESRGAN models remain available in the same modal. See P Image Upscale on NightCafe and Upscaling Your Creations. Clarity Upscaler is a separate creative upscaler that adds detail while upscaling up to 4×.

Original algorithms

NightCafe's earliest algorithms remain available for nostalgia and unique looks until 1 December 2026: Artistic (VQGAN+CLIP, the original text-to-image method), Coherent (CLIP-guided diffusion), and Style Transfer (neural style transfer). They live on the original create forms, not in Studio. See Deprecation of Style Transfer, Artistic, and Coherent.

Retired / deprecated models

Some models have been retired, usually because an external provider discontinued them. These are no longer available for new creations (or are scheduled to be removed on the dates below):

Understanding model attributes

NightCafe rates most image models on four qualities, each scored from 0 to 1 (higher is better). These power the "best for" picks above:

  • Prompt adherence — how faithfully the model follows your prompt.
  • Art — strength at artistic and stylised output.
  • Realism — photorealistic quality.
  • Typography — ability to render readable, correct text.

Base cost is the approximate credits for one standard image. Actual credits depend on resolution, number of images, start/reference images and your plan. See the credits and subscription articles for details.

FAQ

There are too many models — which one should I use? GPT Image 2.5 Flare on Quality Low. It's about 1 credit an image, and it handles generation, editing and reference images. Raise Quality or switch to Sunburst if you want more precision. GPT Image 2.6 is newer and sits higher in the picker, at 2.5 credits for Quality Low. GPT Image 2 Low is still a strong previous-generation alternative.

Which models give the most for my credits? On PRO: GPT Image 2.5 Low (~1), Flux 3 Klein (from 0.75), Muse Image (~2, unlimited on Creator+), Seedream 5.5 (2 at 1K) or Seedream 5.0 Flash (flat 2) and GPT Image 2 Low (~1). Without PRO: Z-Image Turbo (~0.5) and Flux 2 Klein 9B Fast (~0.75). For video: P-Video-2 Fast (~8 for 5s at 720p) is the cheapest current any-PRO clip; MiniMax H3 Max Turbo (~20 for 5s at 720p) is the arena-quality value pick.

Which models are free? Stable Diffusion 1.5 and DreamShaper v8 offer free base generations. Stable Diffusion v1.4 is free too, and it is older and lower quality than 1.5. Most modern models require credits and many require PRO.

Which models can edit my photos? GPT Image 2.5, GPT Image 2 and Muse Image are the strongest everyday options in Studio today. Nano Banana 2.1 (~3.5 at 1K) now ranks near the top of independent editing arenas. Flux 3 Klein is the cheap PRO Flux editor. Seedream 5.5 edits cost 2.75 (1K) or 4.25 (2K). Flash stays a flat 2. Flux 2 Klein 9B Fast or Qwen Image Edit Plus work without PRO. Compare them on the Image Editing tab.

Which models make video? Studio defaults to PixVerse V6. P-Video-2 Fast is the cheapest everyday option on any PRO plan. MiniMax H3 Max and H3 Max Turbo are the arena-quality picks for 5–15 second clips. Seedance 2.5 Flash is the cheaper 480p/720p Seedance. Seedance 2.5 is the 1080p option. See the Video models section — Veo 3.1, Seedance 2.5 Flash, Kling 4.0 Flash, Kling 4.0, Wan 3.0, Grok Imagine Video, Gemini Omni Flash 1.2 and PixVerse are also current. Gemini Omni Flash 1.2 launched on 9 November 2026. Kling 4.0 Flash launched on 5 November 2026. Kling 4.0 launched on 29 October 2026. Gemini Omni Flash (the original) was retired on September 30, 2026. Omni Flash 1.1 shows Gone Soon and leaves on December 3, 2026. Seedance 1.5 Pro was removed on November 11, 2026. GPT Image 1.5 leaves on December 1, 2026.

What happened to Nano Banana (Gemini Flash 2.5)? Google retired it on October 2, 2026. It is no longer available for new creations. Use Nano Banana 2.1 instead. See Deprecation of Nano Banana.

Can I make my own model? Yes — train a LoRA on your own face, object or style. See the LoRA section above.

What does the New badge in the model picker mean? Recently launched models show a New badge for a couple of weeks after they are added. As of 16 November 2026 that includes Seedance 2.5 Flash, Seedream 5.5, GPT Image 2.6, Gemini Omni Flash 1.2, and Kling 4.0 Flash.

What happens to Style Transfer, Artistic, and Coherent? They stop accepting new creations on 1 December 2026. Existing creations stay. See Deprecation of Style Transfer, Artistic, and Coherent.

Model availability and pricing change over time as providers release new models and retire old ones. This guide reflects NightCafe's official Studio lineup and Artificial Analysis arena rankings at the time of writing (most image ranks checked 9 November 2026; Seedream ranks checked 13 November 2026; Seedance 2.5 video rank checked 16 November 2026).