Start with output seconds
H3 pricing is metered by completed output duration, not render time. Ten seconds costs twice as much as five.
Production planning for MiniMax video
Turn an idea into a real production budget. Price one clip, a month of content, or an always-on AI TV channel in seconds.
Published fal pricing at 768p; retry allowance included.
| Model | Resolution | Rate / sec | Estimate |
|---|---|---|---|
| H3 Max | 768p | $0.08 | $92 |
| MiniMax H3 Lowest | 768p | $0.06 | $69 |
Pricing, without the guesswork
Output duration drives the base price. Resolution, retries and the amount of fresh content you generate determine the real monthly number.
H3 pricing is metered by completed output duration, not render time. Ten seconds costs twice as much as five.
Generative video rarely lands on the first attempt. Add a retry allowance to avoid an optimistic production estimate.
For AI TV, a reusable library can cover most airtime. Generate only the percentage that truly needs to be fresh.
H3 field guides
Practical, source-linked notes on H3 Max, ComfyUI, open weights, LoRA, GGUF and local deployment.
Browse all 13 guidesFrequently asked
On fal, H3 Max is listed at $0.05 per output second for 480p and $0.08 per output second for 768p. Promotional rates may temporarily reduce those figures.
At the standard 768p rate of $0.08 per second, a 10-second H3 Max generation costs $0.80 before retries or reference-input charges.
Generated audio is included with the output video rate. Reference-to-video inputs may add separate token-based charges after the included allowance.
H3 Max focuses on fast 480p and 768p generation. Standard H3 also supports 2K and 4K. At 768p, the calculator uses $0.06 per second for H3 and $0.08 per second for H3 Max.
Methodology & sources
Default rates are based on fal's published H3 pricing. Provider prices can change, so the calculator includes a rate override for any quote you want to model.