Max is a different tradeoff, not a universal upgrade.
Evaluate the base H3 route when its output controls and higher-resolution options fit the deliverable. Evaluate Max when iteration speed and prompt adherence are central to the job. Keep the endpoint and resolution attached to every claim.
fal reports speed and quality improvements for its post-trained model. Those are supplier claims; this guide does not convert them into a VSModels benchmark or a promise about your queue time.
Separate the model from the serving configuration.
| Question | MiniMax H3 | H3 Max |
|---|---|---|
| Provenance | MiniMax base model | fal post-training on H3 |
| Shared budget comparison | 480p or 768p | 480p or 768p |
| Higher-resolution price entries | 2K and 4K listed by the endpoint | 1080p listed by the endpoint |
| What to verify | Actual output dimensions and endpoint controls | Actual output dimensions and endpoint controls |
The endpoint lists more resolution choices than some overview copy, which still emphasizes older 2K or 768p limits. Use the current endpoint schema for your intended request. A billing entry alone does not establish native detail or imply that an upscaled output contains more information.
First/last-frame, reference-video and editing workflows may use different routes. Compare those contracts separately; do not infer every input type from the text-to-video page.
Use regular prices after the launch discount.
| Resolution | H3 | H3 Max |
|---|---|---|
| 480p | $0.05 | $0.05 |
| 768p | $0.06 | $0.08 |
| 1080p | No matching entry in this snapshot | $0.16 |
| 2K | $0.13 | No matching entry in this snapshot |
| 4K | $0.16 | No matching entry in this snapshot |
Max’s endpoint advertises a 40% promotion ending October 15, 2026. The table and calculator use its stated regular rates for planning after that promotion. Before checkout, verify which rate applies to your account and selected endpoint.
At 768p, a 10-second attempt is $0.60 on H3 or $0.80 on Max. If H3 gives one accepted result in two billed attempts and Max succeeds first time, the illustrative usable-clip cost is $1.20 versus $0.80. That is a budgeting example, not an observed quality difference.
Compare a batch at the same resolution.
Regular fal text-to-video rates, excluding the Max discount ending October 15. Shared resolutions only; no upscaling, tax, gateway fees or unpaid retry assumption. Acceptance is illustrative, not measured. This calculator does not generate video.
Judge the shot before judging the model.
Motion & identity
Keep the same reference, framing and intended movement. Check subject continuity and object changes across the whole shot.
Sound & instructions
Check the words, speaker, timing and requested silence. Watch with and without sound to separate the errors.
Time & accepted cost
Record submission-to-download time and all billed attempts. A fast rejected clip does not shorten the final edit.
Render at a shared resolution for a quality comparison. Test each model’s higher-resolution workflow separately if that is what your final project needs. This avoids awarding a model a “quality win” simply because one result is displayed larger.
For token-based video budgeting, continue with Seedance 2.5 pricing.
Common questions
Is H3 Max a MiniMax base-model update?
fal describes H3 Max as its post-trained variant of MiniMax H3, optimized with its serving stack. Keep that provenance separate from MiniMax’s base-model release.
Is Max always cheaper?
No. At the inspected regular text-to-video rates, both are $0.05/s at 480p, while H3 is $0.06/s and Max $0.08/s at 768p. A better acceptance rate can still change the cost per usable clip.
Does a fast inference demo measure my wait time?
Not necessarily. Queue time, uploading references, retries and downloading results can affect the time until your usable file arrives. Measure the complete workflow.
Are H3 Max and H3 Max Turbo interchangeable?
No. Treat Turbo as another endpoint with its own settings and pricing. The calculator here compares the ordinary text-to-video routes, not Turbo.
Sources, method and access
Prices and specifications are a dated documentation snapshot. Calculators run in your browser and estimate supplier API costs; they do not call a model or quote VSModels credits. Vendor claims and independent evaluations are labeled separately. VSModels has not run a paid head-to-head test for this guide and does not offer live calls to the models covered here.
- MiniMax · H3 release
- fal · H3 text-to-video endpoint and rates
- fal · H3 Max provenance and reported speed
- fal · H3 Max rates and October 15 promotion expiry
Found a changed rate or a reproducible result? Send the endpoint, settings, date and original source to support@vsmodels.app.