The short answer
MiniMax H3 and H3 Max have the same base video price on SeaDance, the same available durations and the same output resolution. Both include sound and support text, frame-based and still-reference creation. The practical price difference appears when you add several still images in Reference mode. H3 includes more images before an additional reference charge applies. That distinction matters more to a budget than assuming the word Max always means a larger purchase.
Start with MiniMax H3 when your plan calls for several separate visual references and you want to control that expense. Consider H3 Max as another model choice for the same base specification, while checking the quote after attaching references. Neither recommendation is a measured quality ranking. Your own shot requirements should determine which output is usable.
What the names mean
According to MiniMax Research, H3 is designed around joint audio-video generation and multimodal context understanding. According to fal, H3 Max is a post-trained variant of MiniMax H3. These descriptions establish the relationship between the models; they do not establish a SeaDance turnaround guarantee or an independent comparison of image quality. Evaluate a delivered clip against the brief rather than importing a vendor's promotional ranking.
MiniMax's wider model presentation includes capabilities beyond SeaDance's current output tiers. Those wider output settings are not available on SeaDance. The comparison below is deliberately about the purchasable browser workflow: the input mode, accepted still images, duration, sound and credit quote. Keep vendor research demonstrations separate from those controls.
| Setting | MiniMax H3 | MiniMax H3 Max |
|---|---|---|
| Modes | Text to Video; Image to Video; Reference | Text to Video; Image to Video; Reference |
| Durations | 5, 10 or 15 seconds | 5, 10 or 15 seconds |
| Output | 768P; 16:9 only | 768P; 16:9 only |
| Sound | Included | Included |
| Frame inputs | 1–2 images | 1–2 images |
| Reference inputs | 1–9 still images | 1–9 still images |
| Reference images before surcharge | 5 | 2 |
Base prices by duration
These base prices apply to Text to Video and Image to Video, and to Reference requests within the included-image allowance. A frame input does not carry the Reference-mode surcharge. This distinction is useful when the actual task is to animate a prepared opening image: choose the frame workflow for that task rather than uploading the same image as part of a large reference collection.
| Model | Output | 5 seconds: credits / USD | 10 seconds: credits / USD | 15 seconds: credits / USD |
|---|---|---|---|---|
| MiniMax H3 | 768P | 9 credits / $1.81 | 17 credits / $3.41 | 25 credits / $5.02 |
| MiniMax H3 Max | 768P | 9 credits / $1.81 | 17 credits / $3.41 | 25 credits / $5.02 |
Prices checked November 4, 2026. Every dollar estimate uses the Basic monthly plan: $29.90 for 149 credits, or about $0.20 per credit. Dollar amounts allocate a prepaid balance; they are not individual checkout prices. Whole-request credit charges are rounded upward before conversion to dollars. New accounts start with 0 credits and require a paid plan or credit pack to generate.
Reference images change the total
The following examples keep the duration at 5 seconds and the output at 768P. Counts refer to still images in Reference mode. The extra reference calculation happens before the whole request is rounded to an integer credit. As a result, adding an image does not always increase the final integer total immediately. Read the quote for the complete request rather than multiplying a rounded surcharge by the number of images.
| Reference images | H3: credits / USD | H3 Max: credits / USD |
|---|---|---|
| 1 | 9 credits / $1.81 | 9 credits / $1.81 |
| 2 | 9 credits / $1.81 | 9 credits / $1.81 |
| 3 | 9 credits / $1.81 | 10 credits / $2.01 |
| 5 | 9 credits / $1.81 | 13 credits / $2.61 |
| 6 | 9 credits / $1.81 | 14 credits / $2.81 |
| 9 | 12 credits / $2.41 | 19 credits / $3.81 |
Use images with separate jobs
Reference mode is suitable when images describe different aspects of the intended result. A subject photograph can define appearance, another image can establish the environment, and another can suggest the visual treatment. State those roles in ordinary language. Remove an image if it adds conflicting clothing, geometry or lighting without contributing an essential part of the brief. More input material is not automatically a clearer instruction.
The difference between the two reference allowances makes this preparation financially useful. A compact set may fit both models at their base charge; a larger set can lead to very different totals. Select references for relevance rather than filling the upload limit. Keep their order stable while revising the prompt so that instructions such as the product image and the room image continue to refer to the same visual sources.
Use frames for a continuous transition
Image to Video accepts a first frame and an optional last frame for both H3 choices. It is suitable for animating an existing composition or moving toward a prepared endpoint. Keep the subject, viewpoint and environment compatible across those images. Describe the transition between them rather than introducing a new scene that contradicts both. Frame support guides the requested transition; it does not promise a perfectly reproduced final picture.
A product on a tabletop might turn slightly and settle, or a curtain might move while the camera advances gently. A modest continuous action gives you a clear criterion for reviewing the output. If you need several reference roles rather than a defined opening and ending, use Reference mode instead. The frame prompting guide explains that distinction with original examples.
A practical selection process
Begin by choosing the input mode, then the shortest duration that contains the intended action. Check the full quote with the actual reference count. For a base-price comparison, hold the scene and duration constant and look for the required subject details, completed motion and suitable sound. The labels H3 and Max should not replace those checks. A successful output for one composition does not establish a universal winner.
If a take fails creatively, identify whether the problem is the starting composition, conflicting references or an overcomplicated action. Simplifying the brief may be more useful than changing the model immediately. Keep the successful prompt, selected duration and image order together as a reusable recipe for your own project. The cost-per-second comparison helps compare these models with other options without confusing their output sizes.
Frequently asked questions
Is H3 Max always more expensive?
No. The base prices match at every available duration. Larger Reference-mode image sets can make H3 Max more expensive because its included-image allowance is smaller.
Does either model offer a larger output tier here?
Both are sold at 768P in 16:9 on SeaDance. Wider output settings described elsewhere are not available on SeaDance.
Do first and last frames add a reference surcharge?
No. Image to Video frame inputs do not carry this surcharge. The additional-image rules in the table apply to Reference mode.
Why can another reference leave the credit total unchanged?
The underlying amount is calculated before the complete request is rounded upward. An extra reference can remain inside the same rounded credit total.
Does Max guarantee a faster completed request?
No SeaDance turnaround guarantee is established by this comparison. Choose based on the quoted settings and whether the output fits your scene.
Sources
External background checked November 4, 2026. SeaDance settings and credit quotes are described separately above.