MiniMax H3 (Hailuo 03) Review: Resolution, Audio and Price

MiniMax H3 (Hailuo 03): What Actually Shipped

One model, several names. MiniMax calls it H3. Its Hailuo app and a lot of resellers call it Hailuo 03, and you will also see Hailuo 3, Hailuo 3.0 and Hailuo H3. They are the same model, the third generation after Hailuo 01 and Hailuo 02. MiniMax released it on 31 July 2026, and three days later put the weights on Hugging Face. That second move is the one the category will remember.

Hailuo 03 at a Glance: Resolution, Audio and Price

Rechecked against the live API schema and the provider's price list on 12 October 2026:

• Resolution: 480p and 768p are generated natively; 2K (the API default) and 4K are upscaled from a 768p base render. Our Studio offers 768p, 2K and 4K.
• Audio: stereo dialogue, effects and ambience on every clip, with no off switch.
• Length: 5 to 15 seconds per clip, in whole seconds.
• Provider price: 5 cents a second at 480p, 6 at 768p, 13 at 2K and 16 at 4K.
• Studio price: 20 credits for 5 seconds at 768p, 33 at 2K, 40 at 4K; a full 15 seconds at 4K is 120 credits.

The headline specification is straightforward: clips of 5 to 15 seconds, native stereo audio on every generation, and a reference system that takes images, video and audio through the same door. We wired all three of its endpoints into our own Studio the week it landed, which is a different kind of reading than a demo reel gives you. This is what that turned up.

The short version: H3's two genuine advances are that audio is not optional and that references are not fragmented across specialist models. Everything else is competitive rather than category-defining. The open weights matter more than either.

The Audio Is Not a Toggle

Most models in this class treat sound as a feature you enable. Veo has a flag. Seedance has a flag. You decide per generation whether you are paying for a track.

H3 has no flag. We went looking for one, because the integration assumed there would be, and the API simply does not expose it. Every clip returns with a stereo track: dialogue where the prompt implies speech, effects where the action implies impact, room tone underneath.

That cuts both ways and it is worth being clear about which way it cuts for you.

When you want sound, it is excellent. A single pass produces picture and audio that were generated together, so footsteps land on the frame the foot lands on, and a slammed door does not need nudging by three frames in post. Anyone who has synced generated audio to generated picture by hand knows exactly how much time that saves.

When you did not want sound, you pay for it anyway. If your pipeline lays a composed score and a designed effects bed over everything, H3's track is a stem you generate, download and delete. There is no cheaper audio-free tier to fall back to. Price the model accordingly rather than assuming you can opt out.

Omni-Reference Is the Part That Is Genuinely New

The pitch MiniMax makes is that H3 was pretrained to treat reference and editing relationships as language, rather than being assembled from separate expert models for subject reference, motion reference and editing. That sounds like marketing until you use the reference endpoint.

In practice you attach up to nine reference images, up to three reference video clips and up to three reference audio tracks, and then you cite them in the prompt by modality and order. "Image 1" for the character. "Video 2" for the camera move you want echoed. "Audio 1" for the voice. One prompt, one context, one pass.

Compare that with the usual approach, where a character-consistency model, a motion-transfer model and a lip-sync model each own one job and you chain them, losing a little fidelity at every handoff. H3 collapses the chain. On a shot that needs a specific face doing a specific move with a specific voice, that is a meaningfully shorter route.

The limits are real and worth knowing before you brief a shot around them. Reference clips run 2 to 15 seconds each and 15 seconds combined, which is a quarter of what Wan allows. And the audio channel cannot be the only reference you supply. Audio rides along with picture references; it does not drive a generation on its own.

What a Second Costs

At provider list price, H3 is billed per second of output:

Two surcharges sit outside that table and both have caught people out.

Reference images are free for the first five, then 8 cents each. With a nine-image cast that is an extra 32 cents on top of the clip, every time you generate. On a shot you iterate twenty times, the references cost more than several of the renders.

Reference video adds billed seconds. In our Studio a reference clip is charged per second at the same rate as the output, so a 10-second clip you feed in costs the same as 10 seconds of clip you get out. Attach references up to the 15-second combined ceiling and you have doubled the price of a 15-second job before the model has produced a frame. Some models discount reference seconds; we do not on H3.

One more thing the table hides: only 480p and 768p are native. The API schema states that 2K and 4K "upscale a 768P base result". That does not make them worthless, since a clean upscale is still sharper on a large screen, but you are paying 13 or 16 cents a second for resolution the model did not generate. For social delivery, 768p is the honest price of the picture. The 480p tier is new on the provider's side and our Studio does not offer it yet.

For our own credit pricing on H3, see the Key facts box at the top of this page or the Studio rather than any number quoted in an article. Provider rates move, ours follow them, and a figure written into a blog post in August is a liability by October. This table proved it: the provider cut its published 768p rate from 8 cents to 6 after this review first ran. For every model's rate side by side, see what AI video costs per second, and the Render Ledger works out what a finished shot costs once retakes are counted.

MiniMax H3 Max: The Silent Variant

H3 Max arrived on 26 August 2026 as a separate, speed-tuned variant of H3 rather than a resolution tier of it. It shares H3's 5 to 15 second range and aspect ratios, and in our Studio it sits on the same tile. The differences are the reason to pick it:

• No audio at all. Base H3 always renders stereo; Max returns a silent clip, which makes it the H3 to use for B-roll you are going to score yourself.
• A lower ceiling. 480p, 768p or 1080p, where the 1080p tier is a refinement of a native 768p render. No 2K or 4K.
• Text or a start image only. Image-to-video takes an optional end frame, but the Studio runs no reference mode for Max, so the omni-reference workflow above stays with base H3.

At our current rates its 480p tier is the cheapest way into the H3 family: 16 credits for 5 seconds, against 20 for base H3 at 768p.

What Wiring It Up Taught Us

These are the things a spec sheet does not tell you, gathered from actually putting all three endpoints into production.

Duration is an integer, not a string. Veo takes "8s". H3 takes 5 through 15. It is a trivial difference that will silently reject your first three requests while you read the error as something more interesting.

Audio input arrived late, and under an unusual name. At launch, H3's only audio channel was reference_audio_urls on the reference endpoint. We initially exposed an audio-input flag on the assumption there would be a conventional audio_url, and the route happily accepted a supplied track, charged for the job, and then discarded the audio before submitting. By October the text-to-video and image-to-video endpoints had gained a target_audio_url field, which pins a clip of your own (at least 2 seconds) to the soundtrack and replaces the generated audio. Our Studio does not expose it yet. If you are integrating, verify against the live schema rather than the family resemblance to other models, and check it again a month later.

Image-to-video has no aspect-ratio field. The output follows the source image, which is correct behaviour and completely undocumented in the places you would look first. There is a dedicated end-frame slot alongside the start frame, so keyframe transitions work without a separate endpoint.

Check the reference-second ceiling before you quote a price. Without a per-model cap, three 15-second reference clips get quoted and charged as 50 billed seconds before the provider rejects them for exceeding the 15-second combined limit. That is a bill for a job that never ran.

None of these are defects. They are the ordinary friction of a young API, and they are the reason to test a cheap short render on a new endpoint before committing a client deadline to it.

Where It Sits Against the Rest

Against Veo 3.1, H3 wins on duration and on audio being free of a decision, and loses on prompt obedience when the brief is a complicated piece of dialogue direction. Veo did what you asked more often. That comparison has an end date, though: Google is retiring the Veo 3.1 preview models on 22 October 2026, and the Studio replaced Veo 3.1 with Gemini Omni Flash 1.1 on 15 October. Omni shares H3's always-on audio but stops at 10 seconds, and we have not yet set the two against each other on dialogue.

Against Kling, it is not really a contest of the same thing. Kling remains the tool of choice for camera-led motion where you were always going to build the soundtrack yourself, and its motion coherence over a long move is still the reference point.

Against Seedance, H3 is the cheaper route to a spoken shot at 2K. On resolution the two lines now meet at 4K, H3 at its top tier and Seedance 2.0 at its own; on length, Seedance 2.5 goes further, with single takes to 30 seconds against H3's 15.

The honest framing is the one this site keeps arriving at from different directions: these are not substitutes competing for one slot. A working stack keeps three or four and picks per shot. The skill worth developing is knowing which one a given brief wants, not finding the single model that wins everything.

Open Weights Change the Calculation

MiniMax published MiniMaxAI/MiniMax-H3 on Hugging Face on 3 August 2026: a 33-billion-parameter image-text-to-video model. Within days there were ComfyUI builds and Turbo LoRA variants from the community. By 12 October, Hugging Face's own counter showed more than 3.6 million downloads in the previous 30 days.

That is more consequential than any benchmark in the launch post. A frontier-adjacent video model with downloadable weights means the price floor for this capability is no longer set by whoever hosts it. It means fine-tuning on a brand's own footage becomes a thing a studio can do rather than a thing a lab can do. And it means the capability survives the company, which is not nothing in a category where endpoints have quietly disappeared before.

One caution: open weights and open licence are different claims. The model card lists a custom MiniMax H3 Community License Agreement, not Apache or MIT. Read it before you build a commercial pipeline on a local copy, because the download being free tells you nothing about what you are permitted to sell.

Questions People Ask

What is MiniMax H3 (Hailuo 03)? H3, also sold as Hailuo 03 (and listed by some platforms as Hailuo 3 or Hailuo H3), is MiniMax's video generation model, released on 31 July 2026. It produces clips of 5 to 15 seconds with native stereo audio in the same pass, at up to 4K, and it takes text, images, video and audio together as a single reference context rather than through separate specialised models. MiniMax also published the weights on Hugging Face on 3 August 2026.

Does MiniMax H3 generate audio? Always, and you cannot turn it off. There is no generate_audio switch anywhere in the API, unlike Veo where audio is an option you enable. Every H3 clip comes back with a stereo track covering dialogue, effects and ambience. The API can now swap that track for an audio file you supply, but it still bills the same per second, so if you were going to lay your own sound, H3 Max is the silent and cheaper option.

How much does MiniMax H3 cost? At provider list price, H3 is billed by the second: 5 cents per second at 480p, 6 cents at 768p, 13 cents at 2K and 16 cents at 4K. Reference images are free for the first five and 8 cents each after that. In the AIVideos Studio a clip costs 20 credits for 5 seconds at 768p, 33 at 2K and 40 at 4K, and reference video seconds are billed at the same rate as the output.

Is MiniMax H3 open source? The weights are open. MiniMax published MiniMax-H3 on Hugging Face on 3 August 2026, a 33-billion-parameter image-text-to-video model, and the community had ComfyUI builds and LoRA variants running within days. It ships under MiniMax's own H3 Community License Agreement rather than a standard open-source licence, so read those terms before you build a commercial pipeline on a local copy.

Is MiniMax H3 better than Veo 3.1 or Kling? For a self-contained shot that needs sound, H3 is very strong and its 15-second ceiling is longer than most. Veo still has the edge on prompt obedience for complex dialogue direction, and Kling (no longer in our Studio, but still available on its own platform) remains a strong choice for pure camera-led motion where you intend to add audio yourself. They are not really substitutes; the honest answer is that a working stack keeps more than one.

How It Did in Our 48-Render Test

We rendered six shot briefs on eight models at the tiers each one actually supports, and judged every result. MiniMax H3, at 768P and $0.30 a render, delivered none of the six. Richard Byrne's note from judging it: "They weren't terrible, but sometimes it was the output not following the prompt or some physics/realism not being correct." The renders are in our 48-render model comparison and in The Shot Ledger, with the prompt each one received.

Who Should Reach For It

Anyone generating a shot that has to speak. Dialogue, a piece to camera, a character with a voice: one pass, in sync, no manual alignment. This is the case H3 is built for.

Anyone chasing consistency across a sequence. The omni-reference endpoint holds a face, a location and a camera language across shots more cheaply than chaining three specialist tools, provided you respect the 15-second combined reference ceiling.

Anyone who wants the option of running it themselves. The open weights make H3 the only model in this comparison you could, in principle, host, fine-tune and keep.

Not for silent B-roll on base H3, where you pay for an audio track you throw away (H3 Max is the silent option), and not for shots longer than 15 seconds, where you are stitching regardless. If you need a long take with sound, Wan 3.0 reaches 30 seconds in one pass.

You can try H3 through the Hailuo app directly, or generate on it in our Studio alongside the rest of the roster on one credit balance. Worth saying plainly, because the two are easy to confuse: the Hailuo app is MiniMax's own consumer product, while the Studio is our pay-as-you-go console running H3 alongside the rest of the roster. Same model, different front door, different pricing.

AIVideos.eu provides director-level AI video production, custom avatar creation, and commercial brand film workflows. For custom video production enquiries and agency partnerships, contact us through our studio contact form.