Givon AI models

For animating a storyboard, transitions between supplied images and continuing an existing clip. Timed keyframes guide motion through one continuous shot, with optional generated audio.

For staging a continuous shot from a storyboard: key images establish what should appear at selected times. Direct transitions and camera movement, with separate controls for audio and prompt reasoning.

Black Forest Labs' FLUX 3-powered video upscaler enlarges an existing MP4 by 1.5x to 3x while preserving its aspect ratio. Precise mode stays faithful to the source; creative mode restores and invents fine detail with optional prompt guidance.

For short scenes with sound, transitions between supplied frames and reworking an existing clip from written instructions. Visual references guide a new scene; choose another model when silent output is required.

Fast short-form video with synchronized audio and strong prompt following. It can continue from the last frame, making scene stitching easier. 480p/720p.

For short stories, animated images and looping scenes. First, last and timed keyframes guide the progression; references establish subjects and style. Audio can be enabled separately.

For a short clip from a simple idea or animating one image with sound. A simpler Grok 1.5 option for exploring motion and mood without a storyboard of multiple frames.

Alibaba's top video model: produces a clip with synchronized audio and speech in one pass. Use it for cinematic multi-scene videos with prepared voiceover, from text, a frame, references, or source-video edits. 720p/1080p.

For atmospheric scenes, animated illustrations and character clips guided by several images. Choose a wide or vertical composition for the intended platform. Creates a new scene with sound; source-video editing is unavailable here.

Talking avatar from a single photo: the model reads vocal tone and rhythm, then builds lifelike expressions and hand gestures. Sync from text or an existing voiceover.

Synchronized audio in a single pass: speech, ambience, and effects are generated directly in-frame without separate dubbing. Use it for budget clips and talking heads when multi-scene control is not needed.

Affordable motion-control: transfers movement from a video reference to your character. Use it for simpler motion when 3.0-tier precision is not required.

Kling flagship: up to 15 seconds and 4K, stable character identity across scenes, multi-scene direction, and synchronized multilingual audio.

Transfers recorded movement, dance, or gestures from a video sample to your full-body character while locking face identity and capturing complex motion. Use it when choreography fidelity and appearance consistency matter.

For short stories told across several shots with a recurring character, with separate direction for each shot. Images guide appearance; a video guides motion and camera work. Using a video reference requires silent output.

For motion studies and short silent scenes from a brief or a first frame. Turbo is an iteration option in the Kling family; a final frame cannot be specified.

For scenes with a planned sequence of events or several characters. Ordered frames guide the progression; images, videos and voices can guide appearance, movement and sound. Intermediate frames are not pinned to exact times.

Animates a person from a photo and synchronizes speech, natural expressions, and gestures with a voiceover. Useful when you need a speaking or singing presenter from one portrait.

Synchronizes lip movement in an existing video with a new audio track. Use it when the video is already shot and you only need dubbing, localization, or speech replacement.

For silent scenes that combine a first frame with character or product reference images. References guide appearance; strict character locking through Elements is unavailable in our integration.

768P/2K video with stereo audio from text, a first and optional last frame, or image, video and audio references. Up to 9 images, 3 videos and 3 audio clips. Each video is 2–15 seconds, up to 15 seconds combined and 50 MB per file. Videos work without images; audio requires an image or video. Frames cannot be combined with references. Reference-video seconds are billed in addition to output seconds.

H3 Max creates synchronized-audio video from text or frames for 1–15 seconds at 480P/768P/1080P. It supports first/middle/last frames and up to 9 image, 3 video and 3 audio references, with 12 references total. Middle frames require first/last frames, an interior middleFrameTimeSec and native 480P/768P. A soundtrack replaces output audio and remains distinct from the separate LipSync model. Extend adds 1–15 seconds at up to 2K to a 1.625–60 second source; Insert replaces a selected interval; Recast replaces people from 1–4 photos while preserving source motion, cuts and audio. Reframe expands a 0.25–15 second source to a chosen aspect ratio at 480P/768P, preserving source timing and audio; the surroundings prompt is optional.

Animates one image and synchronizes its mouth movement to a supplied voiceover. It supports 480P, 768P, 1080P, and 2K output; the result follows the audio duration up to 14.8 seconds.

H3 Max Turbo creates synchronized-audio video from text, a first frame, a last frame or both for 1–15 seconds at 480P/768P/1080P. An optional soundtrack replaces output audio; LipSync remains a separate model. Extend adds 1–15 new seconds at 480P/768P/1080P/2K to a 1.625–60 second source, with up to three audio references (15 seconds total).

Creates a seamlessly looping GLSL shader from a prompt, image or video, with an MP4 preview and downloadable HTML and source code.

Creates an editable SVG animation with a GSAP timeline from a prompt, image or video, with an MP4 preview and downloadable HTML source.

Talking and singing characters from a single photo: lip movement, expressions and gestures follow the rhythm and delivery of supplied audio. For presenters, promotional messages, educational videos and musical performances.

Ads and short scenes focused on expressive characters and camera movement. Useful for product close-ups with detailed lighting and textures. Create from text or animate a first frame, with optional dialogue, music and sound effects.

Follows director-style commands such as angle, camera motion, and shot changes through text, with audio generated in one pass. Use it for cinematic reference-guided shots up to 4K.

For exploring camera movement, action and reference-guided scene variations. The faster Seedance option for iterations with sound; compare a candidate shot with standard 2.0 before a final take.

For draft checks of motion, composition and a reference setup before a final clip. The entry Seedance option helps explore several directions; available reference types depend on the selected mode.

For stories with several action beats, long continuous takes and scenes built from image, video and audio references. Create a new clip or rework a section of existing footage with direction for action and camera movement.

Changes the background, object, or lighting in existing footage from text, one reference, and an optional mask while preserving the subject, shape, motion, and expressions. Duration comes from the source video; output is 720p or 1080p.

For dubbing, localization and replacing speech in an existing speaker video. Matches lip movement to the selected voiceover and can also animate a speaker photo. The audio recording supplies the words.

For smoother movement and slow motion in an existing clip. Generates intermediate frames to increase frame rate or slow down the action. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

For enlarging AI video, stylized scenes and wide shots by generating texture and detail. A description can guide the new detail. This is creative enhancement: faces, lettering and small objects may change. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

For cleaning up noisy footage, such as low-light video. Reduces noise and compression artifacts, with optional enlargement. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

For cleaning up and enlarging filmed footage. Standard mode offers manual noise and compression controls; Natural mode provides gentler twofold enlargement for a natural texture. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

For restoring and enlarging many clips when turnaround matters. Topaz positions Fast 2 for rapid workflows. Generated detail can change fine features of the source. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

For restoring detail in AI-generated and archival footage: faces, fabric, materials and lettering. Designed for realistic enhancement and reduced detail flicker between frames; review the resulting footage. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

Google's flagship model for premium cinematic shots: up to 4K video with synchronized audio including dialogue, sound effects, and ambience out of the box. Up to 3 references keep character and style stable.

For iterating on advertising and narrative scenes with generated dialogue, sound effects and ambience. Google positions Fast as a quicker Veo option; identical quality to the standard version is not guaranteed.

The most affordable Veo tier: up to 1080p without 4K and audio that can be turned on or off. Use it for high-volume social content when 4K is unnecessary.

Video generation and editing in one engine: from text, from a photo, with a target final frame, or by editing an existing clip from a description. Text and photo generation accept optional input audio. Up to 1080p.

For extended scenes with camera movement and multiple references for characters, objects or action. Images, video and audio guide a new scene. Useful for a long continuous take with sound; it does not edit an existing clip.

For the same reference-guided scenes with sound as Wan 3.0 when turnaround matters. Alibaba positions Prime as the high-speed version; it does not claim a quality advantage over standard Wan 3.0.

For product scenes and visual composites built from multiple source images. Combine objects and a setting in one image, then refine the composition through written edits.

For exploring visual ideas and editing with multiple sources: change a setting, assemble a composition or make variants of a scene. Reference images supply the material for a new image.

OpenAI's previous-generation model for complex images: it follows long instructions, multi-object composition, and multilingual in-frame text with high precision. Use it for infographics, slides, packaging, multilingual posters, and full-image edits in 1K, 2K, or 4K.

For fast iteration, text-to-image generation and reference-guided edits that keep characters and objects recognizable.

For detailed images and precise edits: complex compositions, in-frame text and small elements. Per the developer, it takes longer than Flare.

Base xAI image tier: generate and edit full images from text without masks, and compose from several references. Use it for quick concepts and conversational edits when Pro-level precision is not required.

For stylized illustrations, posters and exploring a distinctive visual direction. Generates and edits images from a brief and uploaded sources. Exact lettering and fine detail need review.

Higher tier of Grok Imagine: more detail, cleaner in-frame text, and stronger composition control from detailed prompts. Use it when the base tier is not sharp enough.

For fashion portraits, editorial shoots and campaign imagery with a defined mood. Preset styles help establish the aesthetic. Creates a new image from text; it does not use uploaded person or product photos.

For poster, cover and illustration concepts with short lettering. Create a layout from scratch or revise a source image with additional references; check wording and fine detail before publishing.

For photorealistic product shots and portraits. Create an image from scratch or edit one source image from a brief, such as changing its setting, clothing or lighting.

The MAI Pro option for staged advertising images and detailed edits of one source. Microsoft positions it for targeted edits, layout changes and lettering; Givon has not established a quality advantage over standard 2.5.

For ad variations, composites and a series featuring the same subject. Combines multiple images into a scene and edits it from written instructions.

Entry tier in Google's image family: the most affordable 1K image generation. Dialog editing and reference blending make it useful for volume work and quick drafts.

Fast and lightweight Nano Banana 2 tier for 1K drafts, high-volume creative iteration, reference-guided ideation, and prompt-based image edits when 2K/4K output is not needed.

Google's updated image generation and editing model: 1K, 2K and 4K output, flexible aspect ratios, and up to ten reference images. Use it for product visuals, text in images and reference-based edits.

Google's premium tier for complex brand scenes, with strong style-guide adherence, reference handling, lighting, and material rendering. Use it for polished product and portrait visuals, multi-object compositions, and 4K finals.

For drafts of menus, infographics, posters and other text-heavy layouts, including edits of an existing image. Check wording, numbers and Russian lettering manually.

For signs, labels and packaging where lettering should look part of the object. The Qwen Pro tier focuses on detailed layouts and lettering, especially in Chinese and English; review the text for accuracy.

For icon sets, illustrations and brand graphics using a preset style. Choose raster images or editable SVG within the same model; style presets establish a shared visual direction for a series.

For logos, icons and flat illustrations intended for further work in a vector editor. Creates editable SVG or raster graphics from a brief. Choose Recraft V4 when you need preset styles.

Cinematic lighting and stable character identity across generations. Use it for product catalogs, character sheets, and reference-guided edits; a reliable workhorse with 2K/4K output and up to 14 references.

Lightweight Seedream 5.0 Lite tier: reasons over complex prompts and can search the web, assembling multi-object scenes and topical visuals. Supports example-based reference edits and output up to 3K.

For ad composition variants and edits guided by several images: use separate references for the subject, objects and setting. Useful for exploring an idea and making successive changes to an existing image.

Seedream for photorealistic commercial hero images, product photography, and edits guided by annotated references. It is strong on natural lighting, skin, and materials while following explicit art direction; 1K/2K output with up to 10 references.

For enlarging AI illustrations and creatively developing texture. Generates new detail from the source content; useful when reinterpretation is welcome and faces, text and objects do not require strict preservation. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

Removes photographic noise at the original image size. Useful for cleaning up grainy photos before further editing. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

Reduces blur and improves photographic sharpness. Choose the processing mode for the source blur, such as camera shake or missed focus. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

For enlarging photographs, CGI and graphics with text while prioritizing source detail. Choose the source type so the processing matches a photo, rendered image or lettering. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

Enlarges logos, cutouts and other transparent images fourfold while retaining transparency in the PNG output. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

For restoring small photographs that lack detail. Reconstructs texture while enlarging the image; useful when reconstruction is acceptable and exact preservation of every feature is not required. Quotes and creation with a spending limit require a previously imported source asset:// URI. Creation without a spending limit also supports secure URL import.

Portrait-first image model: control facial features, makeup, and hairstyle through references. Use it for avatars, beauty assets, and consistent character series up to 2K.

For posters, packaging and complex compositions containing text, tables or formulas. Prompt reasoning works through a detailed brief before generation; review all lettering and numbers in the result.