Blog

Comparison

Gemini Omni 1.2 API on Black Friday: 1080p for $0.0343 a Second

Per-second prices, how they compare with Google and other platforms, and Black Friday math by volume

Nov 16, 2026 3 min read
Gemini Omni 1.2 API Black Friday cover: 1080p at $0.0343 a second
On this page

Gemini Omni 1.2 Flash on HiAPI is $0.0343 a second at 720p and the same at 1080p, and $0.08 a second at 4K. There's no separate charge for input tokens, so a 10-second 1080p clip is $0.343, whatever the prompt length or number of reference images. HiAPI's Black Friday runs Nov 24 to Nov 30 (PT), and Omni counts as video spend for both the top-up bonus and spend-back.

Price per second

Resolution Per second 10-second clip
360p $0.0229 $0.229
720p $0.0343 $0.343
1080p $0.0343 $0.343
4K $0.08 $0.80
  • 720p and 1080p cost the same. If you were going to deliver 1080p, there's no reason to render at 720p to save money.
  • Clips are 3 to 15 seconds, 10 by default, in 16:9 or 9:16. With a first and last frame, the limit is 10 seconds.
  • 1080p and 4K are upscaled from the model's output, and a 4K clip may come back without an audio track. For clips that need sound, 1080p is the safer choice.

How it compares

Per-second rates on each platform's own pricing page, checked on Nov 16 (Beijing time).

Platform 720p 1080p 4K Input tokens
HiAPI $0.0343 $0.0343 $0.08 not charged
Google Gemini API (Standard) about $0.07 about $0.104 about $0.209 $1.50 per 1M
fal $0.0819 $0.1229 $0.2457 $1.50 per 1M
Replicate $0.105 $0.155 $0.31 —
KIE (per clip, 8 s) $0.525 a clip (about $0.066/s) same as 720p $0.945 a clip (about $0.118/s) —
  • Google bills Omni 1.2 by output tokens. Its pricing page converts that into the effective per-second prices above: "approximately $0.07 per second" at 720p, $0.104 for 1080p and $0.209 for 4K.
  • Evolink bills by tokens at the same output rate as Google, so its per-second cost works out about the same as Google's.
  • KIE charges per clip, and its page doesn't name the Omni version.
  • At 360p, HiAPI ($0.0229) is about the same as Google ($0.023). The difference shows up from 720p.

Black Friday by volume

What you make List price What Black Friday gives you Actual cost
300 clips, 10 s, 1080p $102.90 Top up $110, get $8.80 (8%); spend $102.90, get 10% back ($10.29) about $84, 81% of list
1,000 clips, 10 s, 1080p $343 Top up $350, get $35 (10%); spend $343, get 15% back ($51.45) about $257, 75% of list
200 clips, 10 s, 4K $160 Top up $170, get $13.60 (8%); spend $160, get 10% back ($16) about $130, 81% of list

"Actual cost" is what you pay minus what's left in your balance after the event: unused bonus credit plus the spend-back paid on Dec 3. Image, video and audio spend all count toward the spend-back tier. The bonus is capped at $50 per account and spend-back at $100.

At 1,000 clips, that's about $0.26 for a 10-second 1080p clip with sound.

Three ways to spend less

  1. Render 1080p, not 720p. Same price, sharper result.
  2. Write longer prompts if you need to. HiAPI doesn't charge input tokens, so a detailed prompt with up to 10 reference images costs the same as a short one.
  3. Top up once. The bonus is per payment: one $300 top-up gets $30, three $100 top-ups get $24.

New accounts get 200 free credits ($0.10), enough for a 4-second 360p test ($0.0916) before Black Friday.

Model page: Gemini Omni 1.2 Flash. Code: curl and Python examples. More: Omni 1.2 guide · Omni 1.2 vs other video models.

Go to the Black Friday page →

FAQ

Does the per-second price drop during Black Friday? No. Black Friday adds a top-up bonus and spend-back; the list price stays the same, so the effective price per second goes down.

Is the price the same for text-to-video, image-to-video and reference images? Yes. HiAPI prices Omni 1.2 by output resolution and seconds only.

Does 4K include sound? Not always. 4K is upscaled and may come back without an audio track. For clips that need sound, use 1080p, which costs the same as 720p.

Related articles

View all articles →

Start building with persistent AI artifacts

Get your API key and call image, video and audio models through one task-based API.

Read the API docs