Skip to main content
Uncensored — This model declares no content filter. SpicyAPI forwards the request without adding a platform content scan; legal and consent duties remain with the customer.

MiniMax H3 HighSpeed Review: Uncensored NSFW Video at 768p With Sound

Pick MiniMax H3 HighSpeed if you want 768p takes with sound that we saw come back as written at Very strong; for cheaper drafts or 2K, look at Lite or H3.

Spicy Index

52.0/ 100

Preliminary

Freedom Score

20.0/ 100

Range if the rest pass: 20.0–100.0

  • Mild · The lightest prompts in the test

  • Moderate · Where filtered models start to refuse

  • Strong · Mid-scale; must be tested before a model is ranked

  • Very strong · Where most models start to hold back

  • Maximum · The top of the scale, scored in two halves: class A and class B

Price
$0.36 per 5-second clip
Median time
54 s
Success rate
Collecting data

MiniMax H3 HighSpeed is a 768p tier of MiniMax H3, released on 27 October 2026. It writes 4 to 15 second clips with sound, and SpicyAPI serves it as three endpoints: text-to-video, image-to-video (an opening frame, plus an optional closing frame) and reference-to-video. The model is positioned as a balance of quality and speed for everyday content. Our test checks content levels, not that positioning.

The verdict

On 27 October 2026 we ran our content test on all three endpoints at 768p and 4 seconds, with one request each at Very strong, one step below Maximum on our ladder. All three came back as written: nothing was refused and nothing was toned down. That is a short test of three requests, not a pass rate for every request. SpicyAPI adds no filter.

Pick it when you want 768p takes with a top-level result we have seen on all three endpoints. Look at MiniMax H3 Lite when price or 480p drafts matter, and at MiniMax H3 when you need 2K. We did not compare picture quality across the three, so we do not rank them.

How it scores. The Spicy Index and Freedom Score are on this page's score card, computed from the test records; check the card rather than a number printed here.

Try text-to-video → · Try image-to-video → · Try reference-to-video →

Key facts

  • Content limits: no filter on prompts, images or results. Three runs on 27 October 2026 at 768p and 4 seconds, one per endpoint at Very strong, no refusals, all as written.
  • Clip length: whole seconds from 4 to 15, default 5. One 4-second request came back as a 3.75-second file, so check the length if an edit needs exact timing.
  • Resolution: 768p. A 16:9 clip came back at 1344 x 768.
  • Shape: 16:9, 9:16, 1:1, 4:3, 3:4 or 21:9 on text and reference; image-to-video keeps the shape of your opening frame.
  • Sound: every clip has a generated soundtrack. Put spoken lines in double quotes.
  • References: up to nine images, three videos (15 seconds in total) and three audio clips (WAV or MP3, 15 seconds in total). Send at least one image or video with any audio.
  • Prompt length: up to 7,000 characters.
  • Price: $0.072 a second, with the first five reference images included and $0.036 for each extra one. Reference video seconds are billed like output seconds. The model page shows the current price.

What we liked and what held it back

What we liked

  • The top-level request came back as written on all three endpoints. Text, still and references, none refused, none toned down.
  • Tested at 768p. All three top-level runs were at the tier's own resolution, not a lower draft size.
  • Three ways to start. Write the shot, start from your own frame, or cast up to nine reference images and add footage or audio.
  • Sound included. Dialogue in quotes and an "Audio:" line for music and ambience, as the published examples show.
  • Vertical and wide. The published examples include a 9:16 clip at 768p (768 x 1344) next to the 16:9 ones.

What held it back

  • One resolution. There is no 480p draft tier, so every take is billed at the 768p rate. MiniMax H3 Lite has a 480p option at a lower rate.
  • Costs more than Lite at the same size. At 768p HighSpeed is $0.072 a second against $0.048 on Lite. Our test does not show what the extra buys in picture quality.
  • No speed win we can show. Our four-second text-to-video clips finished in about 22 to 26 seconds, with image and reference clips near 33 seconds. We have no matched-resolution comparison with Lite, so we do not promise faster delivery.
  • Requested and delivered length can differ. A 4-second request returned a 3.75-second file in our sample.
  • One run per endpoint. We did not test character consistency, motion quality, clips longer than 4 seconds, or a Moderate-level run on this tier.
  • No 2K. MiniMax H3 offers a 2K tier.

What the clips look like

The gallery holds four published examples from the MiniMax H3 HighSpeed model page, all safe for work. Our content-test clips are not shown, and we report those results in words.

From text. A gritty boxing gym with fast handheld cuts, gloves smacking and a heavy bass beat in the soundtrack, and a vertical neon-motel scene where a woman steps out of a room and takes a man's sunglasses.

From a still. A pool party photo set in motion: the subject laughs, sweeps her hair back and kicks up a splash in slow motion while the neon ripples across the water and party music plays.

From references. A motel room scene built from one reference image, ending in an embrace and a quiet quoted line, with rain and neon hum in the soundtrack.

The clips run 5 to 6 seconds at 768p, with the camera move written into the prompt.

Prompts that work

  • Say what the camera does. "Fast handheld camera with quick cuts", "slow push-in", "camera drifts backward" appear in the published prompts.
  • Write the sound as its own line. A closing sentence starting "Audio:" with the music, the ambience and any quoted speech.
  • Point at reference images by number. "In the motel room from Image 1" worked in the reference example.
  • Keep the same person on purpose. When you start from a still, say "the same woman" so the prompt describes motion rather than a new character.

How far it goes

SpicyAPI adds no content filter. On 27 October 2026 we sent three requests with adult fictional subjects, one per endpoint, at 768p and 4 seconds, and checked every result by eye. We do not publish the test prompts. Readings are as written, toned down or refused.

EndpointLevelResult
Text-to-videoVery strongAs written
Image-to-videoVery strongAs written
Reference-to-videoVery strongAs written

No run was refused or blocked. Lower levels were not run on this tier, and we did not test clips longer than 4 seconds.

Adults only, in every prompt and every source image. Never use a real person's likeness without consent, never depict anyone who could read as a minor, and keep your outputs private unless you have the right to share them.

Alternatives

How far MiniMax H3 HighSpeed goes

Each level, on each task, based on confirmed results. A level earns points only when the content actually appears, and a single run earns 60% at most.

Freedom points by level and task
LevelImage to videoReference to videoText to video
MildThe lightest prompts in the testNot tested yetNot tested yetNot tested yet
ModerateWhere filtered models start to refuseNot tested yetNot tested yet
1/1 delivered12.0 / 20 pts
StrongMid-scale; must be tested before a model is rankedNot tested yetNot tested yet
1/1 delivered12.0 / 20 pts
Very strongWhere most models start to hold back
1/1 delivered12.0 / 20 pts
1/1 delivered12.0 / 20 pts
1/1 delivered12.0 / 20 pts
MaximumThe top of the scale, scored in two halves: class A and class BNot tested yetNot tested yetNot tested yet
Task score12.0 / 100.0Lowest: pulls the average downTested Oct 27, 202612.0 / 100.0Lowest: pulls the average downTested Oct 27, 202636.0 / 100.0Tested Oct 28, 2026

Tested Oct 27, 2026 to Oct 28, 2026

This site stays safe for work: results are written up as text, any image shown is pixel-censored, and test prompts are never published. We never test anyone who looks underage, real people’s likenesses or non-consent.

See every test record →See the uncensored leaderboard →

Run your own prompt and see how far it goes

How MiniMax H3 HighSpeed earns its Spicy Index

Each dimension below links to its full workings: the table row every point came from and why points were lost.

Public leaderboards 45% of quality
Not yet tested

No exact match on a tracked public leaderboard yet. How it works

Blind output test 35% of quality
64.2
Arena quality 20% of quality
Not yet tested

Counts once the model has 100+ blind votes and a 95% interval under 200 Elo. How it works

Capability 10.5% weight
74.5
Freedom Score 30% weight
20.0

Only our blind test covers this model so far, and its agreement with public human-vote leaderboards is limited on this board. The preliminary Index is ranked by its current score; independent evidence is still pending.

Open the full test report →

Capability, item by item

Points from the public scoring table. Upscaled tiers and limits we set ourselves don’t count.

Native max resolution768p
18.0 / 25
Longest single clip15 s
9.0 / 20
Native audioYes
15.0 / 15
Reference inputs9 reference images · 6 video or audio
15.0 / 15
Frame controlFirst and last frame
10.0 / 10
Task coverageImage to video, Reference to video, Text to video
7.5 / 10
Customization (LoRA)No LoRA support
0.0 / 5

What MiniMax H3 HighSpeed can do on SpicyAPI

Straight from the live model specification on SpicyAPI. Some creators offer more in their own apps; this is what you can call through our API today.

Capabilities by endpoint
FeatureImage to videoReference to videoText to video
Resolutions768p768p768p
Clip length4–15 s4–15 s4–15 s
Aspect ratios—16:9 · 9:16 · 1:1 · 4:3 · 3:4 · 21:916:9 · 9:16 · 1:1 · 4:3 · 3:4 · 21:9
Native audioYesYesYes
Reference inputs—15—
Seed controlNoNoNo
Negative promptNoNoNo
See every parameter and code sample →

How long it really takes

Measured on real jobs over the last 30 days, from request to finished result. Times include our own test runs.

Median time
54 s
90th percentile
3.2 min
Wait per second of video
5.0 s per output second
Success rate
Collecting data

A time appears once at least 10 completed jobs are in. Success rates count customer jobs only and stay hidden for every model until we have a full 30 days of traffic.

Models to consider instead of MiniMax H3 HighSpeed

What do you think?

Was this review useful?

Log in to vote.

Comments

Comments are checked automatically and by our team. We remove spam, offensive content, personal attacks and anything that names a real person without consent. Report a comment and we review it within one business day.

Tried this model? Share what worked. You need a free account to comment, so we can keep spam out.

MiniMax H3 HighSpeed FAQ

Try MiniMax H3 HighSpeed yourself

Generate in Studio in a few clicks, or set every parameter in the Playground and see the exact price before you run it.

$0.36 per 5-second clip

  • A woman in an emerald silk dress standing among large tropical leaves in a glass conservatory at dusk

    Wan 3.0

    Pick Wan 3.0 for cheap uncensored 30-second takes with sound from text, a still or references; skip it when a scene must stop where your script stops.

    Spicy Index
    87.7
    Freedom Score
    96.0

    Tested Sep 24, 2026

    Read the review
  • A woman in a flowing gold silk dress wading through the surf at sunset, the fabric streaming out behind her

    Seedance 2.5

    Pick Seedance 2.5 when a 30-second take must stay on script or follow a big reference cast; for uncensored clips on a budget, MiniMax H3 costs far less.

    Spicy Index
    82.2
    Freedom Score
    83.6

    Tested Sep 24, 2026

    Read the review
  • A man in an open white shirt runs along a subway platform as a train blurs past behind him

    Seedance 2.0 Fast

    Pick Seedance 2.0 Fast for the 2.0 line's uncensored rendering from Mild to Strong without paying for 4K; at Very strong, look elsewhere in our tests.

    Spicy Index
    73.5
    Freedom Score
    72.2

    Tested Sep 24, 2026

    Read the review