
MiniMax H3 Max from MiniMax creates cinematic video and supports optional last frame guidance when animating a source image, giving developers more control over how each shot resolves. Atlas Cloud provides pay as you go access through an OpenAI compatible API, with standard pricing of $0.05 per second for H3 Max and $0.025 per second for Turbo. Both variants use transparent per second billing. Start building today.
MiniMax H3 Max 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.
Compare inputs, supported resolutions, video lengths, creative controls, and standard per-second pricing across four MiniMax H3 Max endpoints.
| Modality | Description |
|---|---|
| MiniMax H3 Max T2V API (Text to Video) | Turn a text prompt into a cinematic video at 480P or 768P with a duration from 5 to 15 seconds. Choose 16:9, 9:16, 1:1, or adaptive framing for ads, social content, product visuals, and narrative scenes at the standard price of $0.05 per second. |
| MiniMax H3 Max I2V API (Image to Video) | Starting with a first frame, this endpoint creates a prompt-directed video and can optionally use a last frame for added endpoint control. It supports 480P and 768P output lasting 5 to 15 seconds, making it suitable for animating illustrations, product images, keyframes, and campaign artwork at $0.05 per second. |
| MiniMax H3 Max Turbo T2V API (Text to Video) | For 480P generation at a lower standard price, the Turbo endpoint converts text prompts into cinematic videos lasting 5 to 15 seconds. Its 16:9, 9:16, 1:1, and adaptive aspect ratios support cost-conscious production for previews, social posts, concept tests, and high-volume video workflows at $0.025 per second. |
| MiniMax H3 Max Turbo I2V API (Image to Video) | Animate a first-frame image with a text prompt, then optionally supply a last frame to guide where the sequence finishes. Producing 480P videos from 5 to 15 seconds, this $0.025-per-second endpoint fits rapid image animation, storyboard iteration, visual drafts, and scalable content production. |
MiniMax H3 Max brings text driven video, first and optional last frame animation, 5 to 15 second clips, flexible text video framing, and two price tiers together through one Atlas Cloud API.
MiniMax H3 Max turns a text prompt into a cinematic video at either 480P or 768P. Choose any duration from 5 to 15 seconds, then frame the shot for 16:9, 9:16, 1:1, or adaptive output. This route fits storyboards, social clips, product concepts, and other scenes that begin with words instead of source imagery while preserving clear creative direction.
Start with a first frame image, add an optional last frame, and let MiniMax H3 Max animate the transition under text prompt direction. The standard image route supports 480P or 768P output across 5 to 15 seconds. Use it when an approved product shot, character pose, or opening composition must anchor a short visual sequence with a deliberate finish.
Need a quick draft or a longer beat? Set output length anywhere from 5 to 15 seconds across all four family routes. For text generated clips, select 16:9, 9:16, 1:1, or adaptive framing to match widescreen, vertical, square, or prompt led compositions. These controls keep one endpoint family useful across previews, social placements, and compact narrative moments.
Choose MiniMax H3 Max for 480P or 768P delivery, or switch to the Turbo routes when 480P is sufficient. Standard pricing is $0.05 per generated second for Max and $0.025 per generated second for Turbo. Because both tiers cover text and image inputs from 5 to 15 seconds, developers can align resolution and cost with each production stage.
One Atlas Cloud API key reaches four routes: Max text to video, Max image to video, Turbo text to video, and Turbo image to video. Each route uses per second billing, so costs scale with the clip duration you request. Move between higher resolution Max output and lower cost Turbo generation without rebuilding your access layer, which suits applications that mix polished renders with frequent 480P iterations.
See how MiniMax H3 Max, a leading cross-family competitor, and an earlier MiniMax generation interpret the same two video prompts across cinematic action and stylized storytelling.
An 8 second retro-futurist alpine observatory adventure at midnight: a young observatory apprentice vigorously cranks an enormous brass handle, activating interlocking gears as the radial dome shutters split open beneath a cold indigo sky; a mischievous miniature comet suddenly darts inside, skims across the floor, ricochets around the giant telescope, and whips loose star-chart papers into a glittering vortex. Begin with an extreme low-angle ground-level tracking shot racing directly behind the rolling comet, sparks and icy stardust spraying toward the lens; whip-pan upward across rapidly turning copper-gold gears and synchronized mechanical linkages, then orbit dynamically around the boy as he runs, pivots, and unfurls his vivid vermilion scarf to catch the comet. At the climax, the scarf snaps taut with believable fabric physics—but the captured comet instantly dissolves into a swarm of luminous moths, which spiral upward like a living constellation; the camera tilts and accelerates with them, rushing through the open dome toward the Milky Way. Maintain perfect spatial and character continuity, relentless fluid motion, convincing collisions, fluttering paper, reactive cloth, moving machinery, magical particle transformation, and dramatic parallax throughout. Cold indigo moonlight dominates, contrasted by warm brass machinery, comet glow, and the saturated red scarf; the observatory’s radial dome architecture continually guides the eye toward the action and final ascent. Hand-drawn cel animation blended with exquisitely detailed practical mechanical-model textures, expressive linework, subtle film grain, cinematic fantasy lighting, crisp silhouettes, rich volumetric stardust. Audio: rhythmic brass cranking, accelerating gear clicks, comet whistles and ricochets, snapping fabric, rustling charts, then a crystalline fluttering crescendo as the moths escape into space. No cuts to screens, software interfaces, dashboards, progress bars, charts, captions, subtitles, logos, watermarks, or explanatory text; no slow motion, static filler, broken anatomy, duplicated subjects, discontinuous props, or camera teleportation. 16:9 aspect ratio.
Generated with MiniMax H3 Max 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
An 8 second single-flow micro-narrative titled “Zero-Gravity Ballet Escape on a Salt Lake”: at blue hour, a graceful dancer in a mirror-silver dress races to tighten her pointe-shoe ribbons on a fractured white salt crust; begin with an extreme macro of her fingers pulling the satin knot and the silver toe box scraping crystalline salt, then snap-zoom out as she rises into her first rapid pirouette and the ground violently splits beneath her. Drop into a low-angle tracking shot as she sprints en pointe and performs a fluid sequence of jetés across widening fissures, maintaining perfectly consistent anatomy, face, costume, and momentum while the metallic skirt ripples with realistic weight and wind physics. Orbit tightly around her during the highest leap as a long trailing ribbon is seized by the gale and dynamically transforms into a flock of vivid scarlet origami cranes; the cranes beat their paper wings, catch the ribbon, and suspend her weightlessly for half a second above the collapsing salt before releasing her toward safe ground. Finish with a fast vertical crane rise into a perfectly top-down shot, revealing cobalt-blue cracks and intricate salt patterns framing her silver human silhouette and the circling scarlet cranes. Cold cobalt ambient light, razor-thin silver highlights tracing every fold of fabric, restrained crimson accents, crystalline reflections, airborne salt dust, dramatic depth and scale, hyperreal high-fashion cinematic photography, elegant surrealism, seamless continuous motion, physically convincing cloth, ribbon, paper, wind, jumping, and ground fracture simulation, complex but readable camera choreography, sharp cinematic detail, no slow motion, no static filler, no stop-motion look, no street-documentary realism, no anatomy distortion, no duplicate limbs, no identity or costume drift, no flicker, no cuts that break continuity, no screens, software interfaces, dashboards, progress bars, charts, captions, logos, or on-screen text. Synchronized audio: taut ribbon snaps, pointe shoes striking salt, deep crystalline cracking, rushing wind, paper-wing flutter, and a rising orchestral-electronic pulse that cuts to a brief breathless hush during the suspension, then lands with a sharp percussive impact. 16:9 aspect ratio.
Generated with MiniMax H3 Max 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 campaign concepts and product motion to storyboards, transitions, and platform-ready social clips, MiniMax H3 Max turns prompts or keyframes into adaptable videos that run from 5 to 15 seconds.
Turn written scene briefs into cinematic clips with MiniMax H3 Max across 5 to 15 second durations. Directors and creative teams can preview composition, action, and pacing before committing resources to production.
Animate a product's opening image, then add an optional last frame to guide where the clip finishes. Ecommerce teams can create motion concepts, launch teasers, and visual demonstrations from approved still assets.
Choose landscape, vertical, square, or adaptive framing when generating from text with MiniMax H3 Max. Creators can prepare format-aware concepts for feeds, stories, landing pages, and campaign placements without changing the core brief.
Use text prompts to generate 480P or 768P clips lasting from 5 to 15 seconds. Startup teams can turn product narratives into concise visual proofs for pitches, internal reviews, or early customer presentations.
Begin with one keyframe and optionally set a final frame to shape the clip's visual destination. Designers can explore reveals, before-and-after sequences, and scene transitions while preserving a defined opening composition.
Select the Turbo route for 480P prompt or image animation at the standard $0.025 per second rate. Developers can iterate on short video ideas while keeping usage costs transparent and predictable.
Compare MiniMax H3 Max endpoints with two Atlas Cloud alternatives across input workflow, output resolution, clip length, and last frame support.
| Model | Input Workflow | Output Resolution | Clip Length | Last Frame Input |
|---|---|---|---|---|
| MiniMax H3 Max Text-to-Video | Text prompt | 480P, 768P | 5 to 15 seconds | - |
| MiniMax H3 Max Image-to-Video | First frame and text prompt | 480P, 768P | 5 to 15 seconds | √ (optional) |
| MiniMax H3 Max Turbo Text-to-Video | Text prompt | 480P | 5 to 15 seconds | - |
| MiniMax H3 Max Turbo Image-to-Video | First frame and text prompt | 480P | 5 to 15 seconds | √ (optional) |
| Kling V3.0 Turbo Text-to-Video | Text prompt | 720p, 1080p | 3 to 15 seconds | - |
| Veo 3.1 Lite Start-End Frame to Video | First frame, last frame, and text prompt | 720p, 1080p | 4, 6, or 8 seconds; 1080p requires 8 seconds | √ (required) |
Get started in minutes — follow these simple steps to integrate and deploy models through Atlas Cloud's platform.
Sign up at atlascloud.ai and complete verification. New users receive free credits to explore the platform and test models.
Combining the advanced MiniMax H3 Max models with Atlas Cloud's GPU-accelerated platform provides unmatched performance, scalability, and developer experience.
Low Latency:
GPU-optimized inference for real-time reasoning.
Unified API:
Run MiniMax H3 Max, GPT, Gemini, and DeepSeek with one integration.
Transparent Pricing:
Predictable per-token billing with serverless options.
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 Max is a speed-optimized AI video generation model built on MiniMax H3 and post-trained by fal.ai. On Atlas Cloud, its standard and Turbo endpoints turn text prompts or still images into short videos.
Use text-to-video to create a clip from a written scene description. When a visual starting point matters, image-to-video animates a supplied first frame and can use an optional last frame to guide the ending.
For prompt-only generation, use minimax/h3-max/text-to-video or minimax/h3-max-turbo/text-to-video. To animate an image, choose minimax/h3-max/image-to-video or minimax/h3-max-turbo/image-to-video. Match each request field to the published Playground schema for that endpoint.
Standard endpoints generate videos from 5 to 15 seconds at 480P or 768P. Turbo supports the same duration range at 480P, making resolution one of the main factors when selecting between the two versions.
The text-to-video endpoints support 16:9, 9:16, 1:1, and adaptive aspect ratios. Choose a fixed ratio for a known delivery format, or use adaptive when the composition should determine the framing.
Yes. The image-to-video endpoints animate a required first-frame image and can optionally use a last-frame image. Describe the intended movement and transition in the accompanying text prompt.
Atlas Cloud lists the standard price at $0.05 per generated second for H3 Max and $0.025 per second for H3 Max Turbo, across both generation modes. At these rates, a 10-second output costs $0.50 on standard or $0.25 on Turbo.
Choose the standard version when you need 768P output, since it supports both 480P and 768P at $0.05 per second. Turbo is limited to 480P but lowers the standard rate to $0.025 per second, while retaining text-to-video, image-to-video, and 5-to-15-second generation.
No. Clip duration describes the length of the generated video, while processing time depends on serving conditions and is not specified by the verified Atlas Cloud facts for this page. Measure end-to-end latency before setting application timeouts or throughput expectations.
The four listed Atlas Cloud endpoints accept durations from 5 to 15 seconds per generation. Standard output is capped at 768P and Turbo at 480P, so longer sequences or larger delivery files require downstream assembly or upscaling.
Guides, tutorials, and product updates to help you get the most out of Atlas Cloud.