Every model you need in one place
46 image, video, music, sound and voice models from Google, OpenAI, ByteDance, Kuaishou, Black Forest Labs and more, on one canvas and under one subscription.
Every model is included in your plan. You generate with credits, and no model costs extra.
Every one of them is included in every plan.
Image models
18 modelsNo single model wins at everything, so the picker holds the ones that each win at something: dense on-image text, lifelike product photography, poster detail, logo and layout work.
Plenty Hamon
PlentyLabs
Plenty's default for polished, production-ready results
- Reference images
- Editing
GPT Image 2.5 Sunburst
OpenAI
Premium stills and precise edits
- Reference images
- Editing
Nano Banana 2.1
Google
Higher quality, up to 4K
- Up to 4K
- Reference images
- Editing
Nano Banana Pro
Google
Top-tier quality, thinks before drawing
- Up to 4K
- Reference images
- Editing
GPT Image 2.5 Flare
OpenAI
Fast, sharp everyday images
- Reference images
- Editing
GPT Image 2
OpenAI
For complex prompts and crisp text
- Reference images
- Editing
Nano Banana 2 Lite
Google
Instant results at lowest cost
- Reference images
- Editing
Grok Imagine 2
xAI
Newer Grok, sharper and bolder
- Up to 2K
- Reference images
- Editing
Seedream 5 Pro
ByteDance
Poster-grade detail, pinpoint edits
- Reference images
Seedream 5 Lite
ByteDance
Quick images with web knowledge
- Reference images
FLUX 3
Black Forest Labs
Native 4K, sharp text and layout
- Up to 4K
- Reference images
- Editing
FLUX.2
Black Forest Labs
Lifelike photos, precise edits
- Reference images
- Editing
Seedream 4.5
ByteDance
Rich artistic scenes, accurate spelling
- Reference images
Wan 2.7 Pro
Alibaba
Premium polish for hero images
- Editing
Qwen Image 3
Alibaba
Crisp text, edits from three images
- Reference images
- Editing
Ideogram 4.5
Ideogram
Clean text for posters and logos
- Reference images
- Editing
Recraft V4.1
Recraft
Design-ready images for brand work
GPT Image 1.5
OpenAI
Dependable everyday images
- Reference images
- Editing
Video models
20 modelsText to video, image to video, and reference driven multi-scene generation, from a cheap few-second test to a finished cinematic shot with its own sound.
Seedance 2.5
ByteDance
Our most realistic multi-scene video
- Up to 1080p
- Up to 30s
- Reference images
- Audio
MiniMax H3
MiniMax
2K clips from multimodal references
- Up to 4K
- Up to 15s
- Reference images
Kling 3.0 Pro
Kuaishou
Multi-scene stories with sound
- Up to 15s
- Reference images
- Audio
Kling 3.0 Turbo Pro
Kuaishou
Faster Kling 3.0 at 1080p
- Up to 15s
Wan 3.0
Alibaba
Smoother motion from mixed references
- Up to 1080p
- Up to 30s
- Reference images
- Audio
Wan 3.0 Prime
Alibaba
Faster Wan 3.0, with sound
- Up to 1080p
- Up to 30s
- Reference images
- Audio
Luma Ray 3.2
Luma
Long cinematic shots you direct
- Up to 1080p
- Up to 10s
Grok Imagine Video 1.5
xAI
Animates images fast, with sound
- Up to 1080p
- Up to 15s
- Reference images
- Audio
Vidu Q4
ShengShu
4K clips, consistent characters and voices
- Up to 4K
- Up to 16s
- Reference images
- Audio
MiniMax H3 Max Turbo
MiniMax
Fast video, now up to 1080p
- Up to 1080p
- Up to 15s
MiniMax H3 Max
MiniMax
Fast video with strong prompt adherence
- Up to 1080p
- Up to 15s
- Reference images
Seedance 2.0 Pro
ByteDance
Multi-scene videos from mixed references
- Up to 4K
- Up to 15s
- Reference images
- Audio
Kling o3 Pro
Kuaishou
Control scenes with reference images
- Up to 15s
- Reference images
Veo 3.1
Google
Stunning realism with lifelike sound
- Up to 4K
- Up to 8s
- Reference images
- Audio
Gemini Omni Flash 1.1
Google
Fast results, follows your brief
- Up to 4K
- Up to 10s
- Reference images
- Audio
Kling 2.5 Turbo pro
Kuaishou
Dependable cinematic quality
- Up to 10s
Seedance 1.5 Pro
ByteDance
Polished clips with synced sound
- Up to 1080p
- Up to 12s
- Audio
LTX 2.5 Pro
Lightricks
Polished clips with native sound
- Up to 1080p
- Up to 10s
- Audio
LTX 2.5 Fast
Lightricks
Quick, low-cost clips with sound
- Up to 4K
- Up to 20s
- Audio
PixVerse V6
PixVerse
Social clips up to 15 seconds
- Up to 1080p
- Up to 15s
Music and sound
4 modelsSongs, beds and sound effects, generated on the same canvas as the picture they belong to rather than in another tab.
ElevenLabs Music v2.5
ElevenLabs
Studio-quality songs, sung or instrumental, at the exact length you choose.
- Up to 5 minutes
MiniMax Music 3
MiniMax
Full songs with sung vocals, up to five minutes long.
- Up to 5 minutes
Google Lyria 3.5
Google
Full songs from a written prompt, sung in eight languages.
Mirelo SFX 2.0
Mirelo
Sound effects generated to match what happens in your footage.
Voice
4 modelsVoiceover read straight from the script on your board, with selectable voices and style control.
ElevenLabs v4
ElevenLabs
Expressive voiceover with inline audio tags and exact pronunciation in IPA.
Gemini 3.8 Flash TTS
Google
Expressive voiceover with 30 prebuilt voices and a delivery style you describe.
Gemini 3.1 Flash TTS (Preview)
Google
Fast voiceover with 30 prebuilt voices and plain language style control.
Qwen3 TTS
Alibaba
Text to speech with selectable voices across 11 languages.
Frequently asked questions
Can't find your question?
Contact usEach one is trained differently, so each is best at something different. One puts correctly spelled text into an image, another holds a product's real shape, another moves a camera through a scene convincingly. No single model is best at all of it, which is why the picker holds more than one.
No. Every model on this page is included in your subscription. You spend credits when you generate, the cost is shown before you commit, and no model carries a fee of its own or needs its own contract.
Only if you want to. Every generation surface has a model picker with plain guidance on what each one is good at, and the agent will pick one for you if you would rather describe the job than choose a name.
Yes. Run the same prompt through several models and the results land next to each other on the canvas, at full size, so the choice is something you look at rather than read about.
New models are added to the picker as they ship, and the ones already there stay, so a project halfway through does not change look underneath you. Your products, logos, fonts and references live in your workspace rather than inside a model, and they travel with the work when you switch.
Editable. Whatever made them, images, video, voiceover and music arrive on the canvas as separate layers you can retime, recrop, restyle and rewrite, rather than as a flat file you would have to start over to change.