Pika H3 Flex Pricing: Per-Second Rates, Stems, and the Launch Window
Pika H3 Flex, the post-trained MiniMax H3 model Pika launched on October 29, 2026, bills per second of output video, tiered by resolution. Two things make the bill different from most per-second video APIs: a video reference adds its input seconds to the charge, and exporting the clip's audio as separate stems costs a small per-second surcharge. Everything else — aspect ratio, frame rate, seed control, audio on or off — does not change the price. This page lists the confirmed rates and worked examples. Last checked: 2026-11-03.
The per-second rates
Text-to-video, image-to-video and keyframe jobs, per output second:
| Resolution | Standard | Launch (through Nov 5, 2026) |
|---|---|---|
| 480p | $0.02 | $0.01 |
| 768p | $0.04 | $0.02 |
| 1080p | $0.07 | $0.035 |
| 2K | $0.08 | $0.04 |
| 4K | $0.16 | $0.08 |
1080p, 2K and 4K are not rendered natively: the model drafts at 480p or 768p and Pika's in-job upsampler finishes the same take at the higher resolution — that is why the finished-resolution tier is what you pay.
Reference-to-video with a video input
When the request carries a video reference (for motion, style, or an edit), Pika bills input seconds plus output seconds, at a higher schedule:
| Resolution | Standard, per second (input + output) | Launch (through Nov 5, 2026) |
|---|---|---|
| 480p | $0.06 | $0.03 |
| 768p | $0.12 | $0.06 |
| 1080p | $0.15 | $0.075 |
| 2K | $0.16 | $0.08 |
| 4K | $0.20 | $0.10 |
Image references and audio references are free; only video references move you to this table. Each video or audio reference runs 2 to 15 seconds, video references combined are capped at 15 seconds, and a request takes at most 12 reference assets in total — as many as 9 images, 3 videos and 3 audio files.
Audio stems and the prompt rewriter
- Audio stems: exporting the clip's audio as separate stems (dialogue, score, SFX, foley, ambience) costs an extra $0.001 per generated second, delivered as WAV. A 15-second clip with stems adds $0.015.
- Prompt rewriter: $0.02 per request when you run Pika's prompt rewriter through the API. In Pika's own product it is free.
- Draft-to-video finishes: when you upgrade a 480p or 768p draft to a finished resolution, you pay the difference between the tiers, not the full price again.
Worked examples
- 10-second text-to-video at 768p: 10 × $0.04 = $0.40; at launch rates 10 × $0.02 = $0.20.
- 10-second clip finished at 2K: 10 × $0.08 = $0.80; at launch rates $0.40.
- Reference-to-video at 1080p with an 8-second video reference and 10 seconds of output: (8 + 10) × $0.15 = $2.70; at launch rates (8 + 10) × $0.075 = $1.35.
- Stems on a 15-second clip: 15 × $0.001 = $0.015 on top of the video rate.
Clips run 5 to 15 seconds at 24 fps, in eleven aspect ratios from 21:9 to 9:21, with prompts up to 8,000 characters. None of that changes the bill — duration and resolution tier do.
The launch window: 15 free generations a day, through November 5
Two promotions run from the October 29 launch through November 5, 2026. First, the per-second rates above are halved — the launch column in both tables. Second, Pika API Club members get 15 free H3 Flex generations a day. Fifteen 10-second 768p jobs a day is $6.00 of list value; during the window it costs nothing on top of membership. After November 5 the standard column applies and the daily free tier ends, so re-read Pika's model page before you budget past that date — and treat any per-second figure you read anywhere (including here) as dated.
What Pika H3 Flex costs on Imgo
Nothing, because it is not there: Pika H3 Flex runs through Pika's API, in Pika's own product (from November 2, 2026), or self-hosted from the open weights Pika published on Hugging Face and GitHub. Imgo does not price models it does not run. What you can run today is Seedance 2.0 in Imgo's video generator — up to 15 seconds with native audio and image, video and audio references — where every run quotes its credits live before you submit, on the live model list and Imgo's own pricing. The full comparison of rates, limits and references is on the Pika H3 Flex alternative page, and copyable patterns for both models are in the Pika H3 Flex prompts guide.
Sources: Pika's H3 Flex model page and launch post, checked 2026-11-03. Pika's page is the authority on the current price.
Share
Related Posts
No image
Pika H3 Flex takes prompts up to 8,000 characters, up to 12 reference assets, and returns video with five separable audio stems. Here are five copyable prompt patterns, each naming the mode it runs in and what to check in the result.
No image
Pika H3 Flex bills per output second in five resolution tiers, charges video references by input plus output seconds, and adds a small surcharge for audio stems. Here are the confirmed rates, the halved launch prices, and worked examples. Last checked: 2026-11-03.
No image
Wan 3.0 is billed per generated second in three resolution tiers, with a faster Prime tier at higher rates. Here are the confirmed prices, regional differences, worked examples for 5-, 10- and 30-second clips, and the inputs that do not change the bill. Checked 2026-11-03.