# Sociaro model catalogue

[Browse models and pricing](https://sociaro.com/pricing) · [JSON model index](https://sociaro.com/api/model-index)

## Text

- [Qwen3.7 Max](https://sociaro.com/models/alibaba/qwen3.7-max.md) · `alibaba/qwen3.7-max` — Qwen3.7 Max works through multi-step programming and office tasks. Its million-token context can hold the source code and the project brief together.
- [Qwen3.7 Plus](https://sociaro.com/models/alibaba/qwen3.7-plus.md) · `alibaba/qwen3.7-plus` — Qwen3.7 Plus can answer from long documents and call tools. You can limit how much reasoning it uses on each request.
- [Qwen3.8 Max](https://sociaro.com/models/alibaba/qwen3.8-max.md) · `alibaba/qwen3.8-max` — Qwen3.8 Max can plan and revise software across many steps, keeping requirements, source files and test failures in context.
- [Claude Fable 5.2](https://sociaro.com/models/claude-fable-5-2.md) · `claude-fable-5-2` — Claude Fable 5.2 works through difficult software, mathematics and engineering questions. It can reason about diagrams alongside a written specification.
- [Claude Haiku 5.5](https://sociaro.com/models/claude-haiku-5-5.md) · `claude-haiku-5-5` — Claude Haiku 5.5 handles short, repeatable tasks such as classification and summarization at high volume.
- [Claude Opus 5.5](https://sociaro.com/models/claude-opus-5-5.md) · `claude-opus-5-5` — Claude Opus 5.5 reasons through software migrations and performance problems that span a codebase.
- [Claude Sonnet 5.5](https://sociaro.com/models/claude-sonnet-5-5.md) · `claude-sonnet-5-5` — Claude Sonnet 5.5 handles everyday coding, writing and visual design tasks. It can use a screenshot as context while you refine an interface.
- [GLM-5.2](https://sociaro.com/models/dashscope/glm-5.2.md) · `dashscope/glm-5.2` — GLM-5.2 can work through long engineering tasks using source text and results from tools. A million-token context keeps earlier decisions within reach.
- [Kimi K2.7 Code](https://sociaro.com/models/dashscope/kimi-k2.7-code.md) · `dashscope/kimi-k2.7-code` — Kimi K2.7 Code reasons through software changes and can call development tools. You can cap the reasoning budget for each request.
- [Qwen3.8 Flash](https://sociaro.com/models/dashscope/qwen3.8-flash.md) · `dashscope/qwen3.8-flash` — Qwen3.8 Flash can read a long conversation or codebase in a single request. Use a direct answer for simple questions or allow more reasoning for a difficult one.
- [Qwen3.8 Max](https://sociaro.com/models/dashscope/qwen3.8-max.md) · `dashscope/qwen3.8-max` — This DashScope route serves Qwen3.8 Max. It can keep a long specification in view while building software or drafting a report.
- [Qwen3.8 Omni Flash](https://sociaro.com/models/dashscope/qwen3.8-omni-flash.md) · `dashscope/qwen3.8-omni-flash` — Qwen3.8 Omni Flash accepts text, images, audio and video and responds in writing. It can connect what people say with what happens on screen.
- [GPT-6 Astra](https://sociaro.com/models/gpt-6-astra.md) · `gpt-6-astra` — GPT-6 Astra applies adjustable reasoning to technical problems that combine text and images, including mathematical arguments and engineering designs.
- [GPT-6 Luna](https://sociaro.com/models/gpt-6-luna.md) · `gpt-6-luna` — GPT-6 Luna handles focused, repeatable tasks across large numbers of requests. It can read text or images and return a written answer.
- [GPT-6 Sol](https://sociaro.com/models/gpt-6-sol.md) · `gpt-6-sol` — GPT-6 Sol works on coding tasks across multiple files. Its adjustable reasoning lets you choose how much effort to spend on a tricky change.
- [GPT-6.2 Astra](https://sociaro.com/models/gpt-6.2-astra.md) · `gpt-6.2-astra` — GPT-6.2 Astra builds and revises software from detailed briefs. It also draws on source documents and charts to write reports, slide outlines and analyses.
- [GPT-6.2 Astra Law](https://sociaro.com/models/gpt-6.2-astra-law.md) · `gpt-6.2-astra-law` — Astra for Law pairs GPT-6.2 Astra with a legal search index covering US precedential case law. It can find relevant authorities and compare their reasoning.
- [GPT-6.2 Sol](https://sociaro.com/models/gpt-6.2-sol.md) · `gpt-6.2-sol` — GPT-6.2 Sol handles everyday coding and writing, including tasks that take several rounds of feedback.
- [gpt-oss-120b](https://sociaro.com/models/groq/openai/gpt-oss-120b.md) · `groq/openai/gpt-oss-120b` — OpenAI’s larger gpt-oss reasoning model runs on Groq. Its 131,072-token context can hold long documents, code and calculations.
- [gpt-oss-20b](https://sociaro.com/models/groq/openai/gpt-oss-20b.md) · `groq/openai/gpt-oss-20b` — OpenAI’s smaller gpt-oss reasoning model runs on Groq. It can choose a tool, supply its arguments and use the result in a reply.
- [gpt-oss-safeguard-20b](https://sociaro.com/models/groq/openai/gpt-oss-safeguard-20b.md) · `groq/openai/gpt-oss-safeguard-20b` — This OpenAI model classifies text according to the policy you supply in the prompt. Define your own categories and thresholds for moderation decisions.
- [MiniMax M2](https://sociaro.com/models/minimax/minimax-m2.md) · `minimax/minimax-m2` — MiniMax M2 reasons through changes to code and can call tools to inspect a repository or read test results.
- [MiniMax M2.1](https://sociaro.com/models/minimax/minimax-m2.1.md) · `minimax/minimax-m2.1` — M2.1 follows multi-step development instructions across programming languages, with particular strengths in web and Android projects.
- [MiniMax M2.1 High-Speed](https://sociaro.com/models/minimax/minimax-m2.1-highspeed.md) · `minimax/minimax-m2.1-highspeed` — The High-Speed tier of M2.1 delivers the same model with faster responses for interactive programming and code review.
- [MiniMax M2.5](https://sociaro.com/models/minimax/minimax-m2.5.md) · `minimax/minimax-m2.5` — M2.5 can plan a feature, write code and work through test results. It can also research a question when given search tools.
- [MiniMax M2.5 High-Speed](https://sociaro.com/models/minimax/minimax-m2.5-highspeed.md) · `minimax/minimax-m2.5-highspeed` — M2.5 High-Speed runs the same coding and tool-use model with faster generation for live review sessions.
- [MiniMax M2.7](https://sociaro.com/models/minimax/minimax-m2.7.md) · `minimax/minimax-m2.7` — M2.7 connects clues across logs, traces and source code to help diagnose software problems. It can also revise technical documents over several rounds of feedback.
- [MiniMax M2.7 High-Speed](https://sociaro.com/models/minimax/minimax-m2.7-highspeed.md) · `minimax/minimax-m2.7-highspeed` — M2.7 High-Speed brings the model’s debugging and document-editing skills to live sessions with faster responses.
- [MiniMax M3](https://sociaro.com/models/minimax/minimax-m3.md) · `minimax/minimax-m3` — M3 can reason over up to one million tokens in one request, enough to compare lengthy specifications or trace dependencies across a large codebase.

## Image

- [Qwen Image 2.1 Pro](https://sociaro.com/models/alibaba/qwen-image-2-1-pro.md) · `alibaba/qwen-image-2-1-pro` — Combine as many as ten reference images with a written brief. Qwen can take the subject from one image and place it in the setting or style of another.
- [Qwen Image 3.0](https://sociaro.com/models/alibaba/qwen-image-3.md) · `alibaba/qwen-image-3` — Turn a written scene into a finished image at up to 2K. A negative prompt can rule out distractions, and optional prompt rewriting can fill in an initial sketch of an idea.
- [Qwen Image 3.0 Pro](https://sociaro.com/models/alibaba/qwen-image-3-pro.md) · `alibaba/qwen-image-3-pro` — Create detailed layouts with legible lettering, including Chinese and English text on objects. The Pro model is suited to designs where copy and illustration need to fit together.
- [Wan 2.6 Image](https://sociaro.com/models/alibaba/wan-2-6-image.md) · `alibaba/wan-2-6-image` — Edit or combine up to four reference images from a written instruction. A separate mode can compose text and illustrations together, making it useful for visual explanations.
- [Wan 2.7 Image](https://sociaro.com/models/alibaba/wan-2-7-image.md) · `alibaba/wan-2-7-image` — Generate a scene from text and specify how much of each color should appear. A weighted palette helps carry a campaign’s visual identity into a new 1K or 2K image.
- [Wan 2.7 Image Pro](https://sociaro.com/models/alibaba/wan-2-7-image-pro.md) · `alibaba/wan-2-7-image-pro` — Generate up to 4K artwork from a text brief. The Pro model combines the extra resolution with palette controls for work that needs both fine detail and a defined color scheme.
- [Seedream 4.0](https://sociaro.com/models/bytedance/seedream-4-0.md) · `bytedance/seedream-4-0` — Create a set of related pictures from one brief, or use an existing image to guide a revision. Seedream 4.0 brings both steps into the same workflow.
- [Seedream 4.5](https://sociaro.com/models/bytedance/seedream-4-5.md) · `bytedance/seedream-4-5` — Edit a reference image while preserving the subject’s features and lighting. Seedream 4.5 also improves typography for compositions that put several lines of text beside an image.
- [Seedream 5.5](https://sociaro.com/models/bytedance/seedream-5-5.md) · `bytedance/seedream-5-5` — Build a composition from up to ten references, generate transparent artwork from text, or split a design into layers. Text readback makes it easier to locate lettering for a follow-up edit.
- [Seedream 5.0 Lite](https://sociaro.com/models/bytedance/seedream-5-lite.md) · `bytedance/seedream-5-lite` — Describe how objects, labels and visual elements should relate to each other. Seedream 5.0 Lite can reason through that brief to create a new illustration or revise a reference image.
- [Seedream 5.0 Pro](https://sociaro.com/models/bytedance/seedream-5-pro.md) · `bytedance/seedream-5-pro` — Separate a finished picture into a base image and transparent layers for further editing. Seedream 5.0 Pro can also revise a cutout while keeping its existing alpha channel.
- [GPT Image 2.5 Sunburst](https://sociaro.com/models/gpt-image-2.5-sunburst-2026-09-08.md) · `gpt-image-2.5-sunburst-2026-09-08` — Pin the September 2026 Sunburst snapshot for image workflows that depend on a specific model version. It supports image generation and precise edits with several quality settings.
- [GPT Image 2.6 Flare](https://sociaro.com/models/gpt-image-2.6-flare.md) · `gpt-image-2.6-flare` — Generate or edit images interactively as a design brief changes. Flare can carry a reference product into a new composition while you iterate on the scene.
- [GPT Image 2.6 Sunburst](https://sociaro.com/models/gpt-image-2.6-sunburst.md) · `gpt-image-2.6-sunburst` — Create detailed compositions from text or revise a source image. Sunburst is suited to briefs that specify how multiple objects, textures and details fit together.
- [Ideogram 3.0](https://sociaro.com/models/ideogram/ideogram-3.md) · `ideogram/ideogram-3` — Generate a picture from a prompt and steer its palette and visual style. A saved style code or reference images can carry the same look across a series.
- [Ideogram 3.0 Inpaint](https://sociaro.com/models/ideogram/ideogram-3-inpaint.md) · `ideogram/ideogram-3-inpaint` — Repaint the part of a source image marked in black on a mask. The prompt describes what belongs in that area.
- [Ideogram 3.0 Reframe](https://sociaro.com/models/ideogram/ideogram-3-reframe.md) · `ideogram/ideogram-3-reframe` — Give an existing image a new frame size. Ideogram uses the source image to extend the surroundings when the new frame needs more space.
- [Ideogram 3.0 Remix](https://sociaro.com/models/ideogram/ideogram-3-remix.md) · `ideogram/ideogram-3-remix` — Reinterpret a source image with a prompt. Image weight controls how much of the original composition carries over, and style references can guide the new finish.
- [Ideogram 3.0 Replace Background](https://sociaro.com/models/ideogram/ideogram-3-replace-background.md) · `ideogram/ideogram-3-replace-background` — Keep the subject of a source photo and describe a new setting for it. The model identifies the subject without a hand-drawn mask.
- [Ideogram 3.0 Transparent](https://sociaro.com/models/ideogram/ideogram-3-transparent.md) · `ideogram/ideogram-3-transparent` — Generate an isolated subject as a transparent PNG. Choose the framing and optionally upscale the result by 2× or 4× in the same request.
- [Ideogram 4.0](https://sociaro.com/models/ideogram/ideogram-4.md) · `ideogram/ideogram-4` — Turn a written brief into a 1K or 2K image. Specify lettering, placement and materials in the prompt, and choose a rendering speed for drafts or finished work.
- [Ideogram 4.5](https://sociaro.com/models/ideogram/ideogram-4-5.md) · `ideogram/ideogram-4-5` — Create an image from text or edit the first picture you supply. Additional images can guide appearance, and a mask can limit changes to one area.
- [Ideogram 4.5 Precise Edit](https://sociaro.com/models/ideogram/ideogram-4-5-precise-edit.md) · `ideogram/ideogram-4-5-precise-edit` — Change a specific detail and preserve detail outside the edit. For sources larger than a 2048 × 2048 image by area, a context window can target a region at its original resolution.
- [Ideogram 4.6](https://sociaro.com/models/ideogram/ideogram-4-6.md) · `ideogram/ideogram-4-6` — Generate artwork at up to 4K, or edit a source image using a prompt and visual references. A mask can narrow the change to a specific area.
- [Ideogram 4.6 Precise Edit](https://sociaro.com/models/ideogram/ideogram-4-6-precise-edit.md) · `ideogram/ideogram-4-6-precise-edit` — Make a localized change on a source as large as 8192 × 8192. A context window lets the model work on a small region at full resolution.
- [Ideogram 4.0 Remix](https://sociaro.com/models/ideogram/ideogram-4-remix.md) · `ideogram/ideogram-4-remix` — Transform a single source picture with a prompt. Set the image weight to balance a recognizable subject with a new treatment, then choose the output resolution.
- [Ideogram 4.0 Transparent](https://sociaro.com/models/ideogram/ideogram-4-transparent.md) · `ideogram/ideogram-4-transparent` — Generate artwork directly on a transparent background at a selected output resolution, from 1K to 8K. The alpha channel travels with the PNG.
- [Ideogram Background Remover](https://sociaro.com/models/ideogram/ideogram-remove-background.md) · `ideogram/ideogram-remove-background` — Extract the subject of an existing image as a transparent PNG. One source image is enough to produce a cutout.
- [Ideogram Vector 1](https://sociaro.com/models/ideogram/ideogram-vector.md) · `ideogram/ideogram-vector` — Generate an SVG from a written brief. Up to five source images can guide a new illustration or the redesign of an existing graphic.
- [Ideogram Vectorizer](https://sociaro.com/models/ideogram/ideogram-vectorize.md) · `ideogram/ideogram-vectorize` — Trace a raster image into SVG paths. Choose a simple trace for clean silhouettes or a detailed trace to keep smaller features.
- [Kling Image 4.0](https://sociaro.com/models/kling/kling-4.0-image.md) · `kling/kling-4.0-image` — Create text-rich layouts from a brief and visual references. Kling Image 4.0 can deliver a series of up to ten coordinated images, with resolution options up to 8K.
- [Kling Image 3.0](https://sociaro.com/models/kling/kling-v3.md) · `kling/kling-v3` — Create independent 1K or 2K images from a prompt or a starting picture. Generate several takes on the same brief to compare framing, styling or lighting.
- [Kling Image 3.0 Omni](https://sociaro.com/models/kling/kling-v3-omni.md) · `kling/kling-v3-omni` — Use up to ten reference pictures to guide subjects and settings. Omni can turn those references into a connected image series, with output resolutions up to 4K.
- [P-Image](https://sociaro.com/models/pruna/p-image.md) · `pruna/p-image` — Generate one raster image from text at a chosen aspect ratio or custom size. Bring your own compatible LoRA and adjust its strength across a series.
- [P-Image-Edit](https://sociaro.com/models/pruna/p-image-edit.md) · `pruna/p-image-edit` — Describe a change to the first image and optionally add visual references to guide the edit. Keep its aspect ratio or choose another for the result.
- [Recraft Creative Upscale](https://sociaro.com/models/recraft/recraft-creative-upscale.md) · `recraft/recraft-creative-upscale` — Enlarge a raster image and let the model invent finer detail as it works. Use it when the image can benefit from added texture.
- [Recraft Crisp Upscale](https://sociaro.com/models/recraft/recraft-crisp-upscale.md) · `recraft/recraft-crisp-upscale` — Enlarge and sharpen an existing raster image while retaining the design. The result is available as PNG or lossless WebP.
- [Recraft Remove Background](https://sociaro.com/models/recraft/recraft-remove-background.md) · `recraft/recraft-remove-background` — Isolate a subject from its background and return a transparent raster cutout as PNG or WebP.
- [Recraft V4.1](https://sociaro.com/models/recraft/recraft-v4-1.md) · `recraft/recraft-v4-1` — Create a 1K raster image from a brief. A palette, background color and style references help a series share an art direction.
- [Recraft V4.1 Flash V2](https://sociaro.com/models/recraft/recraft-v4-1-flash-v2.md) · `recraft/recraft-v4-1-flash-v2` — Generate 1K raster images quickly from text. Set the aspect ratio and colors while you explore compositions.
- [Recraft V4.1 Pro](https://sociaro.com/models/recraft/recraft-v4-1-pro.md) · `recraft/recraft-v4-1-pro` — Generate a 2K raster image with room for finer texture and tighter crops. Steer the lighting and materials with a palette or style references.
- [Recraft V4.1 Pro Vector](https://sociaro.com/models/recraft/recraft-v4-1-pro-vector.md) · `recraft/recraft-v4-1-pro-vector` — The Pro vector model draws detailed SVG artwork from a prompt. Supply a palette or a style reference, then edit the shapes in your design tool.
- [Recraft V4.1 Vector](https://sociaro.com/models/recraft/recraft-v4-1-vector.md) · `recraft/recraft-v4-1-vector` — Turn a written brief into editable SVG artwork. Set the colors or supply style references to give related drawings a common look.
- [Recraft Vectorize](https://sociaro.com/models/recraft/recraft-vectorize.md) · `recraft/recraft-vectorize` — Convert an existing raster image into SVG paths. Choose standard tracing or the Pro vectorizer for smoother curves and sharper corners.

## Video

- [HappyHorse 1.1 text to video](https://sociaro.com/models/alibaba/happyhorse-1-1-video.md) · `alibaba/happyhorse-1-1-video` — Turn a written scene into a 3–15 second clip. Choose the framing, from a square crop to a 21:9 panorama, and optionally supply a narration track.
- [HappyHorse 1.1 image to video](https://sociaro.com/models/alibaba/happyhorse-1-1-video-i2v.md) · `alibaba/happyhorse-1-1-video-i2v` — Start a shot from an existing image and describe what should move. The video follows the composition and aspect ratio of the opening frame.
- [HappyHorse 1.1 reference to video](https://sociaro.com/models/alibaba/happyhorse-1-1-video-ref.md) · `alibaba/happyhorse-1-1-video-ref` — Use up to nine reference images to guide the subject and visual style of a new shot. The prompt describes where the subject goes and what it does.
- [HappyHorse 1.0 video edit](https://sociaro.com/models/alibaba/happyhorse-video-edit.md) · `alibaba/happyhorse-video-edit` — Give HappyHorse 1.0 a source clip and describe the change. It renders the edited shot at the length of the source footage.
- [Wan 2.6 image to video Flash](https://sociaro.com/models/alibaba/wan-2-6-i2v-flash.md) · `alibaba/wan-2-6-i2v-flash` — Animate a starting image at 720p or 1080p. Flash can make a single shot or a sequence, with sound when the scene calls for it.
- [Wan 2.7 image to video](https://sociaro.com/models/alibaba/wan-2-7-i2v.md) · `alibaba/wan-2-7-i2v` — Animate an opening image, connect it to a chosen end frame, or continue an existing clip. An optional audio track can drive the performance.
- [Wan 2.7 reference to video](https://sociaro.com/models/alibaba/wan-2-7-r2v.md) · `alibaba/wan-2-7-r2v` — Build a new scene around up to five image or video references. Name the references in the prompt to direct the subjects, setting or voice that the model should carry over.
- [Wan 2.7 text to video](https://sociaro.com/models/alibaba/wan-2-7-t2v.md) · `alibaba/wan-2-7-t2v` — Describe a shot and render it at 720p or 1080p for up to 15 seconds. An audio file can guide the scene when narration or music is already recorded.
- [Wan 2.7 video edit](https://sociaro.com/models/alibaba/wan-2-7-videoedit.md) · `alibaba/wan-2-7-videoedit` — Describe an edit to an existing clip. Supply reference images to guide replacements or change the visual style across the shot, with the option to keep the original sound.
- [Wan 3.0 video](https://sociaro.com/models/alibaba/wan-3-video.md) · `alibaba/wan-3-video` — Wan 3.0 generates up to 30 seconds from a prompt, a pair of frames or a mix of image, video and audio references. Pair a portrait and a speech reference to animate a speaker.
- [Wan 3.0 Prime video](https://sociaro.com/models/alibaba/wan-3-video-prime.md) · `alibaba/wan-3-video-prime` — Prime is the faster Wan 3.0 tier. Create shots up to 30 seconds at 1080p. For a speaking portrait, supply the image and recorded speech together in reference mode.
- [Seedance 2.0](https://sociaro.com/models/bytedance/seedance-2.0.md) · `bytedance/seedance-2.0` — Direct a shot using text, opening and closing images, or visual and audio references. Each input has a specific role, so an image can set the first frame or guide a subject that appears later.
- [Seedance 2.0 Fast](https://sociaro.com/models/bytedance/seedance-2.0-fast.md) · `bytedance/seedance-2.0-fast` — Fast uses the same frame and reference controls as Seedance 2.0 and renders at up to 720p. It is the quicker, lower-cost option for reviewing successive takes.
- [Seedance 2.0 Mini](https://sociaro.com/models/bytedance/seedance-2.0-mini.md) · `bytedance/seedance-2.0-mini` — Mini is the lowest-cost Seedance 2.0 tier, with text, frame and reference inputs and output up to 720p. It makes short motion tests economical.
- [Seedance 2.5](https://sociaro.com/models/bytedance/seedance-2.5.md) · `bytedance/seedance-2.5` — Create clips up to 30 seconds from a prompt or a larger reference collection of up to 30 images, 10 videos and 10 audio tracks. An audio track can also be the only media input.
- [Seedance 2.5 Flash](https://sociaro.com/models/bytedance/seedance-2.5-flash.md) · `bytedance/seedance-2.5-flash` — The faster Seedance 2.5 tier makes clips up to 30 seconds at 480p or 720p. It accepts audio alone or a mix of up to ten image, five video and five audio references.
- [Kling 2.6 image to video](https://sociaro.com/models/kling/kling-2.6-i2v.md) · `kling/kling-2.6-i2v` — Animate a still for five or ten seconds. At 1080p, you can also specify the final image or add native audio.
- [Kling 2.6 text to video](https://sociaro.com/models/kling/kling-2.6-t2v.md) · `kling/kling-2.6-t2v` — Generate a five- or ten-second shot from a written scene in landscape, portrait or square format. Choose 1080p to include native sound.
- [Kling 3.0 image to video](https://sociaro.com/models/kling/kling-3.0-i2v.md) · `kling/kling-3.0-i2v` — Use an opening image and, optionally, a closing image to guide a 3–15 second clip. Choose multi-shot coverage, native sound or output up to 4K.
- [Kling 3.0 Omni video](https://sociaro.com/models/kling/kling-3.0-omni.md) · `kling/kling-3.0-omni` — Work from text, first and last frames, or reference images in one model. Clips run up to 15 seconds, with optional native audio and output up to 4K.
- [Kling 3.0 text to video](https://sociaro.com/models/kling/kling-3.0-t2v.md) · `kling/kling-3.0-t2v` — Write a single scene or describe a sequence of shots. Kling 3.0 renders 3–15 seconds from text with optional native sound, at resolutions up to 4K.
- [Kling 3.0 Turbo image to video](https://sociaro.com/models/kling/kling-3.0-turbo-i2v.md) · `kling/kling-3.0-turbo-i2v` — Start with one image and describe the motion. Turbo renders a 3–15 second clip at 720p or 1080p with sound included.
- [Kling 3.0 Turbo text to video](https://sociaro.com/models/kling/kling-3.0-turbo-t2v.md) · `kling/kling-3.0-turbo-t2v` — Turn a written scene into a 3–15 second audiovisual clip at 720p or 1080p. Turbo includes sound on every render.
- [Kling 4.0 Flash image to video](https://sociaro.com/models/kling/kling-4.0-flash-i2v.md) · `kling/kling-4.0-flash-i2v` — Animate a single opening frame in the fast Kling 4.0 tier. A prompt can direct the action over 3–20 seconds at 720p.
- [Kling 4.0 Flash Omni video](https://sociaro.com/models/kling/kling-4.0-flash-omni.md) · `kling/kling-4.0-flash-omni` — Combine images, video references and reusable elements, or edit a supplied clip. The Flash tier renders 3–20 seconds at 720p.
- [Kling 4.0 Flash text to video](https://sociaro.com/models/kling/kling-4.0-flash-t2v.md) · `kling/kling-4.0-flash-t2v` — Render a written scene in the fast Kling 4.0 tier. Each clip is 3–20 seconds at 720p.
- [Kling 4.0 image to video](https://sociaro.com/models/kling/kling-4.0-i2v.md) · `kling/kling-4.0-i2v` — Direct up to 30 seconds of motion with as many as ten keyframes. Opening, middle and closing images can mark the beats of a shot at 720p or 1080p.
- [Kling 4.0 Omni video](https://sociaro.com/models/kling/kling-4.0-omni.md) · `kling/kling-4.0-omni` — Direct a new shot or edit footage using keyframes, reference images, video, reusable elements and voices. The model supports up to 30 seconds at 720p or 1080p.
- [Kling 4.0 text to video](https://sociaro.com/models/kling/kling-4.0-t2v.md) · `kling/kling-4.0-t2v` — Write the action, camera direction and sound into one prompt. Kling 4.0 renders 3–30 seconds at 720p or 1080p.
- [MiniMax H3](https://sociaro.com/models/minimax/h3.md) · `minimax/h3` — Generate a 4–15 second clip with sound from text, opening and closing frames, or image, video and audio references. The standard H3 tier offers 768p and 2K output.
- [MiniMax H3 Highspeed](https://sociaro.com/models/minimax/h3-highspeed.md) · `minimax/h3-highspeed` — Highspeed prioritizes turnaround for 4–15 second H3 clips with sound at 480p or 768p. It accepts the same prompt, frame and reference modes as the standard model.
- [MiniMax H3 Lite](https://sociaro.com/models/minimax/h3-lite.md) · `minimax/h3-lite` — Lite is the lowest-cost H3 tier. It generates 4–15 second clips with sound from prompts, frames or references at 480p or 768p.
- [MiniMax H3 Max](https://sociaro.com/models/minimax/h3-max.md) · `minimax/h3-max` — H3 Max generates 5–15 second shots with sound at 480p or 768p. Direct it with text, frames or a set of references.
- [MiniMax H3 Regeneration](https://sociaro.com/models/minimax/h3-regeneration.md) · `minimax/h3-regeneration` — Render one of your existing 768p MiniMax H3 jobs again at 2K. The source job supplies the take, its duration and its sound.
- [PixVerse C1](https://sociaro.com/models/pixverse/c1.md) · `pixverse/c1` — Describe multiple shots in a prompt and C1 cuts between them. It can start from text, animate an image or connect a pair of frames, with optional generated audio.
- [PixVerse V6](https://sociaro.com/models/pixverse/v6.md) · `pixverse/v6` — Generate from text, animate one image or connect a first and last frame. V6 makes clips from 1–15 seconds, with optional sound and output up to 1080p.
- [PixVerse Video Expand](https://sociaro.com/models/pixverse/video-expand.md) · `pixverse/video-expand` — Change the aspect ratio of an existing clip by generating picture beyond its edges. The model works with source videos up to 30 seconds long.
- [P-Video](https://sociaro.com/models/pruna/p-video.md) · `pruna/p-video` — Generate motion from a prompt or an image at 720p or 1080p. A lower-cost draft mode lets you test the shot before rendering the finished version.
- [P-Video-2-Pro Avatar](https://sociaro.com/models/pruna/p-video-2-pro-avatar.md) · `pruna/p-video-2-pro-avatar` — Animate a portrait using a recorded voice or a written script. The model matches mouth movement and facial expression to the speech and supports videos up to three minutes.
- [P-Video-2-Pro Reference](https://sociaro.com/models/pruna/p-video-2-pro-reference.md) · `pruna/p-video-2-pro-reference` — Direct a 5–15 second clip with sound using a prompt and at least one reference image or video. You can also add audio to guide the result.
- [P-Video-2-Pro Replace](https://sociaro.com/models/pruna/p-video-2-pro-replace.md) · `pruna/p-video-2-pro-replace` — Replace one or two people in a source clip using their reference portraits. The edit retains the original camera, action, background and soundtrack.
- [P-Video-Upscale](https://sociaro.com/models/pruna/p-video-upscale.md) · `pruna/p-video-upscale` — Increase an existing video’s resolution up to 4K and optionally set its frame rate from 16 to 60 fps. Turbo provides a lower-cost processing mode.

## Audio

- [Eleven Multilingual v2 text to speech](https://sociaro.com/models/elevenlabs/eleven_multilingual_v2.md) · `elevenlabs/eleven_multilingual_v2` — ElevenLabs Multilingual v2 reads long passages aloud in 29 languages with a steady delivery. Each request can contain up to 10,000 characters.
- [Eleven v3 text to speech](https://sociaro.com/models/elevenlabs/eleven_v3.md) · `elevenlabs/eleven_v3` — Eleven v3 turns a script into expressive speech in over 70 languages. Its changes in emphasis and emotion suit character performances and dramatic narration.
- [Scribe v1 speech to text](https://sociaro.com/models/elevenlabs/scribe_v1.md) · `elevenlabs/scribe_v1` — Scribe v1 transcribes speech from an audio or video recording in over 90 languages. It can detect the language or use one you specify.
- [Whisper large v3 speech to text](https://sociaro.com/models/groq/whisper-large-v3.md) · `groq/whisper-large-v3` — OpenAI’s Whisper large v3 transcribes multilingual speech on Groq. It handles recordings with background noise, varied accents or technical vocabulary.
- [Whisper large v3 Turbo speech to text](https://sociaro.com/models/groq/whisper-large-v3-turbo.md) · `groq/whisper-large-v3-turbo` — Whisper large v3 Turbo runs on Groq and prioritizes transcription speed. It turns multilingual recordings into text in the original language.
- [Sonilo video to music](https://sociaro.com/models/sonilo/video-to-music.md) · `sonilo/video-to-music` — Sonilo composes original music to match the pacing of an edited video, up to six minutes long. You can request alternate scores or a mix that dips the music under dialogue.
- [Sonilo video to sound effects](https://sociaro.com/models/sonilo/video-to-sfx.md) · `sonilo/video-to-sfx` — Sonilo generates a sound-effects track timed to the action in a video, up to eight minutes long. A prompt can guide the sound of a particular scene.
