Model Guides
Use these guides when you need model-specific workflow IDs, supported inputs, output limits, pricing behavior, or prompting guidance. Availability can vary by Sogni interface and account; the live model catalog in the app or API remains authoritative.
Premium image and GIF
GPT Image 2.6 Sunburst and Flare
Images, animated GIFs, planning complexity, and annotation-guided edits. Open-weight image MOSS-Image
Uncensored 50-step and 8-step generation, English or Chinese captions, layouts, and Personal LoRAs. Open-weight image FLUX 3 Image Klein
Text-to-image, image-to-image, and edits from up to ten optional reference images in one model. Mature-theme image Seedream 5.5 Uncensored
ByteDance Seedream 5.5 on a more permissive content policy: text-to-image, edits from up to 10 reference images, transparent backgrounds and layer extraction, at up to about 4.6 megapixels. Mature-theme image Seedream 5.0 Pro Uncensored
ByteDance Seedream 5.0 Pro on a more permissive content policy: text-to-image and edits from up to 10 reference images, up to 2048 × 2048, and layer extraction into transparent PNGs. Native video LTX-2.5
Fast, HQ, and Pro video workflows with native synchronized audio. Video upscaling FlashVSR v1.1
Promptless upscaling of a finished video to 1080p or 1440p that keeps every frame, the frame rate, and the audio. Video upscaling RelayVSR
Promptless upscaling of a finished video to 1080p, 1440p, or 4K that keeps every frame, the frame rate, and the audio. Open-weight video MiniMax H3 & H3 Turbo
Text, frame-guided, and multi-reference video workflows, with FastH3 Turbo up to 6× faster than standard. Premium video Seedance 2.5
Long-form multimodal video, editing, and extension workflows, real people after a one-time likeness agreement, plus the Uncensored versions of Seedance 2.5 and 2.0 Mini for mature themes. Premium video Wan 3
Unified text, frame-guided, reference, and native-audio generation. Mature-theme video Wan 3 Uncensored
The Wan 3 Enhanced Team's more permissive Wan 3 route, with native audio and mixed media references. Object selection SAM 3
Select an object with points, boxes, or text and get a lossless mask or cutout at the source size. Video object selection SAM 3.1 Video
Select one object on a frame and track it through a clip, returning a lossless mask at the original dimensions and timing.
Images, animated GIFs, planning complexity, and annotation-guided edits. Open-weight image MOSS-Image
Uncensored 50-step and 8-step generation, English or Chinese captions, layouts, and Personal LoRAs. Open-weight image FLUX 3 Image Klein
Text-to-image, image-to-image, and edits from up to ten optional reference images in one model. Mature-theme image Seedream 5.5 Uncensored
ByteDance Seedream 5.5 on a more permissive content policy: text-to-image, edits from up to 10 reference images, transparent backgrounds and layer extraction, at up to about 4.6 megapixels. Mature-theme image Seedream 5.0 Pro Uncensored
ByteDance Seedream 5.0 Pro on a more permissive content policy: text-to-image and edits from up to 10 reference images, up to 2048 × 2048, and layer extraction into transparent PNGs. Native video LTX-2.5
Fast, HQ, and Pro video workflows with native synchronized audio. Video upscaling FlashVSR v1.1
Promptless upscaling of a finished video to 1080p or 1440p that keeps every frame, the frame rate, and the audio. Video upscaling RelayVSR
Promptless upscaling of a finished video to 1080p, 1440p, or 4K that keeps every frame, the frame rate, and the audio. Open-weight video MiniMax H3 & H3 Turbo
Text, frame-guided, and multi-reference video workflows, with FastH3 Turbo up to 6× faster than standard. Premium video Seedance 2.5
Long-form multimodal video, editing, and extension workflows, real people after a one-time likeness agreement, plus the Uncensored versions of Seedance 2.5 and 2.0 Mini for mature themes. Premium video Wan 3
Unified text, frame-guided, reference, and native-audio generation. Mature-theme video Wan 3 Uncensored
The Wan 3 Enhanced Team's more permissive Wan 3 route, with native audio and mixed media references. Object selection SAM 3
Select an object with points, boxes, or text and get a lossless mask or cutout at the source size. Video object selection SAM 3.1 Video
Select one object on a frame and track it through a clip, returning a lossless mask at the original dimensions and timing.
For platform-wide access rules and subscription coverage, see Plans & Pricing and Sogni Unlimited Plans. Developers can retrieve the live catalog through the applicable Sogni API or SDK model-listing surface.