AI Models
Active model catalog
Syntina isn't built on a single model. It brings together different model families within the same product for generating a scene from scratch, editing an existing image, and turning an image into video.
Category Overview
Model families
Right now, a total of 31 models are listed on the public side, grouped under three main generation categories. If you're generating a brand-new scene, look at Text to Image; if you're transforming an existing render or reference, look at Image to Image; if you're animating a single frame, look at Image to Video.
| Category | Model count | What it's used for |
|---|---|---|
| Text to Image | 5 | Models used to generate a brand-new scene, mood, presentation image or concept from a prompt. |
| Image to Image | 15 | Models used to transform an existing image, revise it, change its style, or increase its resolution. |
| Image to Video | 11 | Models used to turn a still image into a moving video, add camera motion, or generate a video between a first and last frame. |
Text to Image
Text to Image models
Models used to generate a brand-new scene, mood, presentation image or concept from a prompt.
| Model | Provider | What it's good for | Quick note |
|---|---|---|---|
| GPT Image 2 | OpenAI | Strong prompt adherence and high scene consistency for new image generation. | Photorealistic, High detail |
| Nano Banana Pro | For workflows that need a strong balance of speed and quality. | Fast, General purpose | |
| FLUX 2 Flex | Black Forest Labs | For more controlled composition and typography-focused generations. | Controlled, Design-focused |
| WAN 2.6 Text | WAN | A practical option for generating fast variations and reviewing your first round of ideas. | Variation, Quick test |
| Grok Imagine | xAI | Flexible generation for creative visual composition and concept exploration. | Creative, Concept |
Image to Image
Image to Image models
Models used to transform an existing image, revise it, change its style, or increase its resolution.
| Model | Provider | What it's good for | Quick note |
|---|---|---|---|
| GPT Image 2 Edit | OpenAI | A powerful edit model for masked editing and controlled scene revisions. | Edit, Inpainting |
| Nano Banana Pro Edit | For teams that want fast image editing with strong quality. | Fast edit, High quality | |
| Nano Banana 2 Edit | Interprets multiple references together to perform semantic editing. | Multi-reference, Semantic edit | |
| FLUX 2 Flex Edit | Black Forest Labs | Stands out in revisions that need tighter control over style and composition. | Style control, Composition |
| FLUX 2 Turbo Edit | Black Forest Labs | Optimized for fast iteration and generating frequent variations. | Turbo, Fast iteration |
| FLUX 2 Edit | Black Forest Labs | For general-purpose visual transformation workflows that need balanced quality and speed. | Balanced, General use |
| WAN 2.6 | WAN | A lightweight alternative for generating fast transformations from a reference image. | Fast, Variation |
| Qwen Image Edit | Alibaba | Strong in image editing scenarios guided by a clear prompt. | Prompt-driven, Edit |
| Seedream 4.5 Edit | ByteDance | Delivers strong editing for architectural images while preserving scene structure. | Architectural revision, High quality |
| Seedream v5 Lite Edit | ByteDance | The faster-running member of the Seedream family, suited to day-to-day iteration. | Fast, Seedream |
| Topaz Upscale | Topaz Labs | Used to deliver an existing image at a higher resolution. | Upscale, Detail enhancement |
| Magnific Creative | Magnific | For creative generations that want a more stylized and interpretive upscale. | Creative upscale, Stylization |
| Magnific Precision v2 | Magnific | Focuses on preserving detail in photorealistic renders and architectural images. | Precision, Photorealistic |
| Longcat Image Edit | Longcat | Preserves longer context to produce in-scene editing and variation. | Long context, Edit |
| Grok Imagine Edit | xAI | Delivers flexible results on the creative transformation and visual interpretation side. | Creative edit, Concept |
Image to Video
Image to Video models
Models used to turn a still image into a moving video, add camera motion, or generate a video between a first and last frame.
| Model | Provider | What it's good for | Quick note |
|---|---|---|---|
| Veo 3.1 | Google DeepMind | For workflows that need high quality, scene depth, and support for generating audio. | Premium video, Audio support |
| Veo 3.1 First-Last | Google DeepMind | Creates smooth transition videos between a first frame and a last frame. | First-last, Transition video |
| Veo 3.1 First-Last Pro | Google DeepMind | Produces the first-last frame flow at a higher quality tier. | Pro quality, First-last |
| Seedance 1.5 Pro | ByteDance | A strong option for presentation videos with fast, high-quality motion generation. | Fast motion, Presentation video |
| Kling 2.5 Turbo Pro | Kuaishou | For speed-focused video generation and quick iteration rounds. | Turbo, Quick test |
| Kling 2.6 Pro | Kuaishou | An improved version with audio, first-frame support, and more controlled motion. | Audio, Controlled motion |
| Kling 3.0 Pro | Kuaishou | For more cinematic motion and higher-end video quality expectations. | Cinematic, High quality |
| Kling O1 Standard | Kuaishou | Appeals to workflows looking for balanced quality and speed within the current Kling family. | Balanced, General video |
| Kandinsky 5 Pro | Sber | A good fit for short video tests and more experimental motion trials. | Short video, Experimental |
| Sora 2 Pro | OpenAI | For projects that need higher quality and longer-duration video generation. | Pro, Long video |
| Happy Horse 1.1New | Alibaba | Generates synced-audio, multilingual lip-synced video up to 1080p from a single image; stands out in talking-head scenarios. | Synced audio, Lip-sync, Talking-head |
Choosing a model directly affects credit consumption too. That's why the natural next stops from this page are the credit system and pricing pages.