MiniMax H3 Max brings Max and Max Turbo into one family for 5–15-second videos at 480p, 768p or 1080p. Both variants create scenes from text or animate a starting image with an optional ending frame; Max also combines subject images, reference video and reference audio. Describe the action, camera and audible events to develop performances, transitions and scene concepts. Choose the task by its accepted sources, output settings and current rate, then review visual continuity and generated sound in the actual result.
MiniMax H3 Max is developed by MiniMax. Vidgo provides access to the model. Model names and trademarks belong to their respective owners.For text-only scenes or frame animation, compare the Max and Max Turbo task rates at the same planned duration and resolution, then test your own brief. Choose Max Reference to Video for at least one subject image with optional movement and sound sources. All five cards open their existing tasks, where you can configure the exact inputs and inspect the complete estimate.
MiniMax H3 Max Turbo Text to Video turns text into coherent 5–15 second videos with synchronized native stereo audio and fast generation. It follows physical motion while rendering subtle changes in light and shadow.
MiniMax H3 Max Turbo Image to Video turns still images into 5–15 second videos with start and end frame control and synchronized native stereo audio. It preserves the source subject and textures while adding realistic physical motion.
MiniMax H3 Max Text to Video turns text prompts into cinematic video with native stereo audio, faster-than-real-time generation, and precise instruction following. It renders complex camera work and physical motion with fine detail across multiple resolutions and common aspect ratios.
MiniMax H3 Max Image to Video turns still images into video with native stereo audio, start and end frame control, and lip-synced dialogue. It preserves the source subject's identity and textures while adding fluid, natural motion and camera work.
MiniMax H3 Max Reference to Video creates cinematic video with combined image, video, and audio guidance. It preserves character appearance and motion rhythm while adapting the aspect ratio and delivering multiple HD quality levels.
Frame animation uses one starting image followed by an optional ending image. Max reference generation uses one to nine images and optional video or audio, with up to twelve assets combined; assign the sources appearance, scene, movement and sound roles in the prompt. Use each workflow’s own input fields so the chosen composition and source material match the intended shot.
| Model / task | Input and workflow |
|---|---|
| MiniMax H3 Max Turbo Text to Video | MiniMax H3 Max Turbo Text to Video turns text into coherent 5–15 second videos with synchronized native stereo audio and fast generation. It follows physical motion while rendering subtle changes in light and shadow. |
| MiniMax H3 Max Turbo Image to Video | MiniMax H3 Max Turbo Image to Video turns still images into 5–15 second videos with start and end frame control and synchronized native stereo audio. It preserves the source subject and textures while adding realistic physical motion. |
| MiniMax H3 Max Text to Video | MiniMax H3 Max Text to Video turns text prompts into cinematic video with native stereo audio, faster-than-real-time generation, and precise instruction following. It renders complex camera work and physical motion with fine detail across multiple resolutions and common aspect ratios. |
| MiniMax H3 Max Image to Video | MiniMax H3 Max Image to Video turns still images into video with native stereo audio, start and end frame control, and lip-synced dialogue. It preserves the source subject's identity and textures while adding fluid, natural motion and camera work. |
| MiniMax H3 Max Reference to Video | MiniMax H3 Max Reference to Video creates cinematic video with combined image, video, and audio guidance. It preserves character appearance and motion rhythm while adapting the aspect ratio and delivering multiple HD quality levels. |
The existing examples include glacier dialogue, a fresco reveal, a taiko performance, wheelchair basketball and a BMX hop. Each record has its own prompt and sources. Inspect their action progression, camera movement and sound to understand the controls used, then choose output duration and resolution for your intended use.
Use Max Text to Video to specify a subject, environment, restrained action and camera path, including desired speech or ambience. The glacier record requests a head turn, brief English dialogue and cave sounds in one shot, providing concrete visual and sound details to evaluate.
Provide one starting frame and an optional ending frame, then describe the intervening movement and details to retain. The Max fresco example uses a covering sheet and a revealed artwork to guide a continuous lowering action while preserving the restorer, chapel and wall design.
Select Max Reference to Video and assign each source a role. The taiko request uses an image for the performer, video for the courtyard, drum and camera, and audio for the attack and resonance. Review identity, geometry and the strike’s timing against the full input set.
Describe an ordered activity, physical contact and sound in the Turbo text task, then choose duration, resolution and framing. The basketball record follows one bounce, a chair turn and a pass in a continuous shot; inspect the result for wheel rotation, hand contact and audible timing.
Supply ordered first and last images to Turbo Image to Video and describe a manageable motion between them. The BMX request specifies a short hop, smooth landing, stable bicycle geometry and tire sounds, making endpoint alignment and physical contact useful review points.
The Turbo BMX case provides an opening and landing frame, then asks for a single controlled hop with hands and feet held on the bicycle. The Max taiko case provides the performer’s image, a courtyard motion clip and a drum-audio source, then asks for one decisive strike and its resonance. Read the full prompts and inspect all inputs with their own results; each demonstrates a separate task and creative brief.
Create a single continuous five-second live-action shot connecting these exact first and last frames. The same helmeted adult BMX rider in the plain orange hoodie makes a small controlled hop from the shallow bank lip, keeps both hands on the handlebars and both feet on the pedals, and lands both wheels smoothly on the flat. Preserve the bicycle frame, wheels, helmet, face and skatepark geometry. Follow the rider with a gentle low tracking camera, without cuts, flips or spins. End at the supplied final pose and composition. Realistic short flight arc, wheel rotation, tire compression on landing and settling cloth. Native rolling tires, one soft landing sound and a satisfied breath, no music. No brands, captions or on-screen text.
Use Image 1 for the adult female taiko performer's exact face, tied-back hair and indigo happi costume. Use Video 1 for the temple courtyard, large drum and the restrained left-to-right camera movement. Use Audio 1 for the dry wooden attack and low resonant drum atmosphere. In one continuous five-second cinematic shot, she raises both sticks, makes one decisive strike on the large drum, and holds the finish as her sleeve settles. Preserve her identity and drum geometry throughout. Match the drum impact to the generated sound, with the resonance spreading through the quiet courtyard. Late golden sunlight, subtle determined expression, no crowd or dialogue, no text, subtitles, logos or advertisements.
Start with a written scene, approved frame images or references for identity, movement and sound. Describe a clear action and the scene details to preserve, then generate with Max or Turbo using the appropriate input method. Reviewed clips can support dialogue studies, artwork reveals, physical-action rehearsal, sports concepts, reference-led performances and atmosphere proposals.
Describe an adult character, a short spoken line and the gesture that accompanies it. Review expression, mouth movement and audible delivery in the generated clip for a narrative scene draft.
Provide starting and ending images showing a covered and revealed object. Describe the physical action and lighting to retain, then inspect continuity for an exhibition, restoration or storyboard concept.
Prepare an action brief or an athlete’s starting frame and describe one concise movement sequence. Inspect body mechanics, equipment shape and contact in the generated video before developing the scene further.
Use compatible first and last frames for a small jump, turn or landing. Describe the path and final pose, then compare both endpoints and movement for a sports-storyboard transition.
Prepare a performer image, appropriate movement video and sound source that meet the reference requirements. Assign each role and review identity, setting and sound timing for a performance concept.
Describe a setting, subtle environmental movement and the sound that should accompany it. Generate at the intended resolution and inspect how light, material and ambience combine for a scene-development proposal.
Text and frame tasks use the selected Max or Turbo resolution rate multiplied by output seconds. Max reference generation starts with the same duration-based video charge, then adds server-metered reference usage above the included 4,096 reference tokens at the configured rate per 1,000 tokens. The reference task displays a minimum video amount while actual media usage is measured by the server. Use the dynamic table for current starting rates and units, and review the complete request estimate or final usage with the selected task.
| Model | Provider | Workflow | Input material | Output specification | Starting rate |
|---|---|---|---|---|---|
| MiniMax H3 Max Turbo Text to Video | MiniMax | Text to Video | Text prompt | 1080p | $0.025 / second |
| MiniMax H3 Max Turbo Image to Video | MiniMax | Image to Video | 1–2 starting/ending frames | 1080p | $0.025 / second |
| MiniMax H3 Max Text to Video | MiniMax | Text to Video | Text prompt | 1080p | $0.05 / second |
| MiniMax H3 Max Image to Video | MiniMax | Image to Video | 1–2 starting/ending frames | 1080p | $0.05 / second |
| MiniMax H3 Max Reference to Video | MiniMax | Reference to Video | 1–9 images; optional video/audio | 1080p | $0.05 / second |
Test your creative brief online, then bring the verified request into your application.
Open the selected task from the model list above. Describe the action, camera and audible events, provide the required frames or references, and set duration and resolution. Review the estimate, noting that reference usage is measured on the server, then compare the generated clip with every source for identity, physical motion, endpoint framing and sound.
Choose a task, inspect its settings and cost, and follow the result from an online test to an authenticated API request.
For text or first/end-frame animation, choose Max or Max Turbo and compare the current rate at your intended resolution. Choose Max Reference to Video when a subject image should combine with optional video and audio cues. Set a whole-number duration from 5 to 15 seconds and 480p, 768p or 1080p; prepare reference media according to that task’s technical requirements.
Open the selected task from the model list above. Describe the action, camera and audible events, provide the required frames or references, and set duration and resolution. Review the estimate, noting that reference usage is measured on the server, then compare the generated clip with every source for identity, physical motion, endpoint framing and sound.
Compare Max and Turbo using the planned resolution and output duration. For Max reference generation, budget for the video minimum plus reference tokens above the included allowance; the server measures the actual media usage. Use the full amount and intended number of tests when planning the project.
Read the actual prompt and inspect every source image, video or audio alongside its generated result. Use the recorded examples to understand how to describe input roles, action progression and the details to preserve in your own request.
Retain the task ID returned by the API to query status and retrieve completed file URLs. Use your online generation history to review earlier results and the creative direction you tested.
Use the documentation linked from each task for its exact model identifier, accepted inputs, authentication and response format. Integrate the same task you tested online so its media and settings match the request.
MiniMax H3 Max is a Vidgo video family containing three Max tasks and two Max Turbo tasks. Both variants support text and first/end-frame generation at 480p/768p/1080p for 5–15 seconds. Max additionally supports image-led references with optional video and audio, technical media requirements and separately metered reference usage.
Explore model introductions, API guides and platform comparisons to understand input preparation, request integration and cost planning for your project.
