Seedance 2.0 Mini & Fast API at Lowest Prices Worldwide — up to 68% off official pricing
MiniMax H3 Fast Subject Preserving Video

MiniMax H3 Fast Subject Preserving Video

MiniMax H3 Fast is MiniMax’s prompt-driven video generation family for cinematic text-to-video, controllable first-frame animation with an optional last frame, and reference-based clips that retain the source subject. Access every workflow through Atlas Cloud’s unified API at the standard pay-as-you-go rate of $0.046 per second, with one integration across the family. Start building today.

MiniMax H3 Fast is developed by MiniMax. Atlas Cloud (operated by Atlas Cloud AI LLC) provides access to it and does not own it. All trademarks belong to their respective owners.

MiniMax H3 Fast Inputs and Video Generation Modes

Compare MiniMax H3 Fast endpoints by source input, verified capabilities, and intended production workflow.

ModalityDescription
MiniMax H3 Fast T2V API (Text to Video)Turn a text prompt into a cinematic 480P video lasting 5 to 15 seconds. Choose 16:9, 9:16, 1:1, or adaptive framing for concept clips, social content, and visual storyboards.
MiniMax H3 Fast I2V API (Image to Video)Starting from a first frame image, this endpoint creates a 480P video guided by text and can optionally incorporate a last frame. It suits image animation, shot transitions, and 5 to 15 second product or character clips.
MiniMax H3 Fast R2V API (Reference to Video)A reference image anchors the subject while a text prompt directs the resulting 480P video. Use this mode for subject consistent promotional shots, creator content, or short visual sequences lasting 5 to 15 seconds.

MiniMax H3 Fast Across Three Creative Paths

MiniMax H3 Fast brings text driven creation, first and optional last frame animation, and reference guided subject continuity into 480P clips lasting 5 to 15 seconds, with flexible framing and one Atlas Cloud API priced at the standard $0.046 per generated second.

MiniMax H3 Fast Text to Video

Turn a text prompt into a cinematic 480P clip with MiniMax H3 Fast. Choose a duration from 5 to 15 seconds and frame the result in 16:9, 9:16, 1:1, or adaptive format. Describe the subject, action, camera movement, and setting in one request. This mode suits direct concept exploration, social video drafts, and shot planning before production.

First and Last Frame Control

Begin with a supplied first frame and add an optional last frame to guide how the sequence opens and resolves. A text prompt directs the movement between those visual anchors while output remains 480P and 5 to 15 seconds long. The input image establishes the starting composition. Use this route for controlled transitions, animated key art, and product reveals with a defined finish.

MiniMax H3 Fast Reference Continuity

Keep a subject from a reference image recognizable while a text prompt places it into motion. MiniMax H3 Fast Reference to Video produces 480P clips from 5 to 15 seconds, giving developers a dedicated route for reference driven generation. Rather than redesigning the subject for every take, carry its visual identity into a fresh scene. It fits recurring characters, mascots, and product centered sequences.

Flexible Five to Fifteen Second Clips

Select any supported duration from 5 through 15 seconds to match the pace of the idea. Shorter clips keep rapid iteration focused, while longer outputs leave room for a complete visual beat within one generation. The same duration range is available across text, image, and reference driven routes. Build compact ads, single shot narratives, or motion tests without changing model families.

Formats for Every Feed and Canvas

Shape text generated videos for 16:9 landscapes, 9:16 vertical feeds, or 1:1 square placements, with adaptive framing also available. That choice lets one model family cover widescreen presentations, mobile social posts, and balanced square creative. Image driven generation uses the supplied first frame as its visual starting point. Pick the route and framing that match the destination instead of rebuilding the concept around one fixed canvas.

One API with Per Second Pricing

Access all three MiniMax H3 Fast routes through Atlas Cloud instead of integrating separate providers for text, image, and reference workflows. Standard pricing is $0.046 per generated second, so cost scales directly with selected clip length. The shared access path simplifies switching among generation modes as a project evolves. It is a practical fit for developers testing concepts or operating repeatable video pipelines.

MiniMax H3 Fast in a Three Model Prompt Showdown

See how MiniMax H3 Fast, a leading competitor, and another MiniMax H3 model interpret the same two cinematic video prompts.

Prompt

A 7-second high-energy micro-story in a storm-threatened fishing harbor: a lithe, weathered silver-haired fisherwoman in a rust-red oilskin coat, mustard knit cap, dark waders, and sea-worn boots chases a runaway fishing net that violent wind has inflated into a gigantic sail, dragging bright orange and yellow floats toward the churning sea. Open with an extreme macro shot of a soaked rope knot snapping, fibers exploding toward the lens, followed by a rapid push-in; whip-cut to a ground-skimming lateral tracking shot as she sprints across wet planks, ducks fluidly beneath a chaotic maze of whipping ropes, steps onto a violently rocking wooden rack, and leaps to wrap both arms around the main cable. Cut to a perfectly vertical top-down shot: the net’s geometric grid, her twisting body, crossing ropes, and bouncing colored floats collide in a clear kinetic composition as her full body weight yanks the cable down and the sail-like net crashes onto the dock with a thunderous wet slap, spraying mist and seawater. Rapidly push into her relieved, exhilarated face as a tiny silver fish springs from the collapsed mesh and lands neatly inside her knit cap; she freezes, crosses her eyes upward, then grins. Maintain exact character, wardrobe, net, rope, and harbor continuity across every cut; physically accurate rope tension, knots, fabric deformation, wind pressure, inertia, collisions, wet surfaces, sea spray, and hair movement. Cold cyan storm light, rust-red coat and orange-yellow floats, dramatic backlight outlining airborne mist, realistic 35mm maritime cinema, handheld urgency, shallow depth of field, subtle film grain, crisp natural motion, fast comic timing, no slow motion. Audio: rising gale, rigging creaks, ropes snapping and whipping, boots hammering wet boards, wooden rack groaning, net slamming heavily, then a tiny comic fish plop and her breathless chuckle. No screens, software interfaces, dashboards, progress bars, charts, captions, subtitles, logos, watermarks, explanatory text, duplicate people, character drift, wardrobe changes, anatomy errors, tangled or melting limbs, impossible rope behavior, weightless cloth, jumpy motion, frozen action, or disconnected cuts. 16:9 aspect ratio.

Generated with MiniMax H3 Fast Text-to-Video on Atlas Cloud

Generated with Kling V3.0 Turbo Text-to-Video on Atlas Cloud

Generated with MiniMax H3 Text-to-Video on Atlas Cloud

Prompt

An 8-second photorealistic cinematic food-commercial sequence set in a dim-sum kitchen at dawn: a young pastry chef lifts the lid of a bamboo steamer, and a glossy, semi-translucent har gow suddenly springs out, its pink shrimp filling glowing through the delicate ivory wrapper. Begin from inside the steamer in the dumpling’s POV as the lid opens and warm golden window light pierces swirling steam; cut to an extreme macro tracking shot skimming along the flour-dusted worktop as the dumpling tumbles, rebounds, and squashes elastically while fleeing, kicking up fine flour with every impact. The chef reacts instantly and sweeps a wooden rolling pin across its path; whip-pan into a rotating overhead shot as the dumpling veers between utensils, compresses against the counter, then launches over the polished spine of a cleaver in one continuous, physically convincing arc. Follow tightly at counter level with a fast diagonal composition as flour particles collide with trailing steam; the dumpling lands on the steamer rim, wobbles, flips back inside, and mischievously blasts a final puff of steam into the chef’s surprised face. Maintain perfect character, dumpling, kitchen, and spatial continuity across every cut; nonstop fluid action, realistic momentum, collisions, elastic deformation, translucent moist wrapper texture, appetizing shrimp detail, tactile bamboo grain, shallow depth of field, crisp macro focus pulls, ivory white, bamboo brown, and shrimp pink palette, warm golden morning backlight, premium high-speed food advertising cinematography, playful suspense and precise comic timing. Synchronized audio: bamboo-lid clack, soft rubbery bounces, floury skids, rolling-pin whoosh, metallic cleaver ping, and a punchy steam hiss, with light rhythmic percussion building to the final gag. No slow motion, no static filler, no screens, software interfaces, dashboards, progress bars, charts, captions, text, logos, watermarks, extra limbs, duplicate objects, broken utensils, food morphing, or discontinuous motion. 16:9 aspect ratio.

Generated with MiniMax H3 Fast Text-to-Video on Atlas Cloud

Generated with Kling V3.0 Turbo Text-to-Video on Atlas Cloud

Generated with MiniMax H3 Text-to-Video on Atlas Cloud

From Prompt to Short Video with MiniMax H3 Fast

MiniMax H3 Fast turns text prompts, opening frames, optional ending frames, and subject references into 480P videos for storyboards, social formats, guided transitions, character shorts, and campaign concept tests.

MiniMax H3 Fast Storyboards

Turn a text prompt into a 480P cinematic clip lasting 5 to 15 seconds. Filmmakers and creative teams can test scenes, pacing, and visual directions before committing to larger production workflows.

Social Campaign Variants

Choose 16:9, 9:16, 1:1, or adaptive framing when generating video from text. Marketing teams can prepare landscape, vertical, and square clips for campaign concepts without rebuilding each idea from a separate starting asset.

Still Image Animation

Animate a first-frame image into a 480P clip guided by a text prompt. Designers can bring illustrations, product stills, or campaign key art to life for presentations, previews, and social posts.

Guided Frame Transitions

Provide both opening and ending frames to guide how an image-based clip develops. This workflow suits transition studies, before-and-after concepts, and storyboard beats that need a defined visual destination on screen.

MiniMax H3 Fast Subject Shorts

Keep a chosen subject present by generating from a reference image and text prompt. Creators can produce character-focused shorts, product appearances, or recurring campaign moments that begin from the same visual subject.

MiniMax H3 Fast Concept Tests

Set each generation between 5 and 15 seconds at 480P for focused idea exploration. Small studios can compare prompts, opening images, and reference subjects across compact clips before selecting directions to develop.

MiniMax H3 Fast in a Four Model Video API Comparison

Compare MiniMax H3 Fast with three Atlas Cloud text-to-video endpoints across clip length, available resolutions, aspect ratios, and listed standard price.

ModelOutput DurationAvailable ResolutionsAspect RatiosStandard Price
MiniMax H3 Fast Text-to-Video5-15s480P16:9, 9:16, 1:1, adaptive$0.046/second
Seedance 2.0 Mini Text-to-Video4-15s or automatic480p, 720p, 720p-SR, 1080p-SR, 1440p-SR16:9, 4:3, 1:1, 3:4, 9:16, 21:9, adaptive$0.056
Wan-3.0 Text-to-video2-30s or smart duration480P, 720P, 1080P16:9, 9:16, 4:3, 3:4, 1:1, adaptive$0.05
Veo 3.1 Lite Text-to-video4s, 6s, or 8s720p, 1080p16:9, 9:16$0.05

How to Use MiniMax H3 Fast on Atlas Cloud

Get started in minutes — follow these simple steps to integrate and deploy models through Atlas Cloud's platform.

Create an Atlas Cloud Account

Sign up at atlascloud.ai and complete verification. New users receive free credits to explore the platform and test models.

Why Use MiniMax H3 Fast on Atlas Cloud

Combining the advanced MiniMax H3 Fast models with Atlas Cloud's GPU-accelerated platform provides unmatched performance, scalability, and developer experience.

Performance & flexibility

Low Latency:
GPU-optimized inference for real-time reasoning.

Unified API:
Run MiniMax H3 Fast, GPT, Gemini, and DeepSeek with one integration.

Transparent Pricing:
Predictable per-token billing with serverless options.

Enterprise & Scale

Developer Experience:
SDKs, analytics, fine-tuning tools, and templates.

Reliability:
99.99% uptime, RBAC, and compliance-ready logging.

Security & Compliance:
SOC 2 Type II, HIPAA alignment, data sovereignty in US.

MiniMax H3 Fast API FAQs

MiniMax H3 Fast is a video generation model on Atlas Cloud with separate text-to-video, image-to-video, and reference-to-video routes. The verified Atlas configuration produces 480P clips lasting 5 to 15 seconds.

Create a cinematic clip from a text prompt, animate a first-frame image with an optional last frame, or generate from a reference image while keeping the subject recognizable. Every verified workflow supports 480P video lasting 5 to 15 seconds.

Select the Atlas Cloud route matching your input, then submit the prompt and media fields defined in its Playground schema. Use minimax/h3-fast/text-to-video, minimax/h3-fast/image-to-video, or minimax/h3-fast/reference-to-video as the model ID.

All three Atlas Cloud routes support 480P output and durations from 5 to 15 seconds. Requests outside this resolution or duration range are not part of the verified configuration.

For text-to-video, choose 16:9, 9:16, 1:1, or adaptive. The verified facts do not establish the selectable aspect-ratio list for the image-driven routes, so follow each route's Playground schema.

Yes. The image-to-video route animates a supplied first frame and can accept an optional last frame, while your text prompt directs the motion between them.

Choose reference-to-video when a reference image should anchor the subject across the generated clip. Pair the image with a prompt describing the desired scene and motion, then review the result because generative consistency is not absolute.

Atlas Cloud lists a standard price of $0.046 per generated second for its text-to-video, image-to-video, and reference-to-video routes. The final charge follows output duration and uses the standard price rather than a temporary discounted rate.

First confirm that the model ID matches the intended workflow, the output is set to 480P, and the duration is between 5 and 15 seconds. For text-to-video, also use 16:9, 9:16, 1:1, or adaptive, then compare every field with the corresponding Atlas Cloud Playground schema.

Explore More Families

Seedance 2.5

Seedance 2.5 API is now available on Atlas Cloud! It gives developers ByteDance's newest video model. It generates up to 30 seconds of native video in a single pass from text, a single image, or as many as 50 multimodal references, with synchronized audio and in-frame multilingual text. On Atlas Cloud you reach it through one key, with subject consistency and improved physics keeping long shots coherent. (Update: Seedance 2.5 1080P API Is Available NOW!)

View Family

Wan 3.0

Wan 3.0 API is the next generation of Alibaba's Wan video family, built to push long-form generation, multi-reference control, and audiovisual quality to new heights. Atlas Cloud already hosts Wan 2.7, 2.6, and 2.5, and Wan 3.0 runs on the same unified key with no separate setup. Start building today. Scroll down to the showcase to see what Wan 3.0 can create.

View Family

MiniMax H3

MiniMax H3 is MiniMax's multimodal video family for text, image, and reference guided creation. Across supported routes, it preserves subjects from reference media, offers flexible aspect ratios, and pairs generated sound with visuals through H3 Developer, with output profiles selected by endpoint. Atlas Cloud unifies the family behind one OpenAI-compatible key with transparent pay-as-you-go pricing from the standard rate of $0.038 per second. Start building today.

View Family

Seedream 5.0 Pro

Seedream 5.0 Pro API gives developers ByteDance's controllable image editing model on Atlas Cloud. It places edits precisely with anchors and coordinates, separates images into editable layers, fuses multiple references, and matches exact colors and materials, with multilingual text at 2K and 3K. On Atlas Cloud you reach it through one key!

View Family

Seedance 2.0

Seedance 2.0 is ByteDance’s production video model for precise shot creation. Turn prompts into video, animate a first-frame image with optional last-frame guidance, or shape results with reference media and optional web search. Atlas Cloud brings these workflows into one unified API with transparent pay-as-you-go pricing and one OpenAI-compatible key. Start building today.

View Family

GPT Image 2.5

The gpt-image-2.5 family from OpenAI gives developers a choice of Flare and Sunburst for production image workflows. Render at arbitrary resolutions up to 3840x2160 and select from five quality tiers, including xhigh and max, to match specific output requirements. Atlas Cloud provides ready-to-use REST inference with no cold starts and standard pricing from $0.004 per generation. Start building today.

View Family

GPT Image 2

The GPT Image 2 API gives developers access to OpenAI's latest image model, the successor to GPT Image 1.5. It generates and edits images with accurate text rendering across Latin and CJK scripts, plus strong composition for posters, mockups, and infographics. On Atlas Cloud you reach it through one unified API alongside 300+ models, with free credits, 99.99% uptime, and no OpenAI organization verification required.

View Family

Gemini Omni Flash

The gemini omni API brings Google DeepMind's natively multimodal Gemini Omni Flash family, including Gemini Omni 1.1 Flash, to developers. Create cinematic video with synchronized native audio, animate still images with precise start and end frame control, or revise existing footage through text guided edits that preserve untouched content. Atlas Cloud provides one OpenAI-compatible key, unified access, and transparent pay-as-you-go pricing. Start building today.

View Family

Grok Imagine

The Grok Imagine API covers xAI's image, video, and speech models, from Image 2.0 to Video 1.5 and xAI TTS v1. Render 1K or 2K stills across 14 aspect ratios, push a scene to 15 seconds of 1080p motion, steer shots with up to 7 reference images, or narrate them in 20 languages. Atlas Cloud runs every mode on one endpoint, priced pay-as-you-go from $0.02 per image and $0.05 per second. Start building today.

View Family

Google

Google's most powerful creative models are all available on Atlas Cloud. Veo 3.1 delivers cinematic video generation, Nano Banana 2 powers high-fidelity image creation, and Gemini brings multimodal intelligence to every workflow. Access the full Google model suite through one API key with Day-0 availability and pay-as-you-go pricing.

View Family

Seedance 2.0 Mini

The Seedance 2.0 Mini API is the lightest, lowest-cost tier of ByteDance's Seedance video line, built for teams where throughput and unit cost matter more than maximum polish. Use it for batch generation, rapid prototyping, and draft passes, all through one OpenAI-compatible key on Atlas Cloud.

View Family

ByteDance

From cinematic video generation to high-fidelity image creation, ByteDance's most powerful models are live on Atlas Cloud. Run Seedance and Seedream at scale with the lowest inference pricing and zero infrastructure overhead.

View Family

One API for All Media AI.

Explore all models