7 product reels: Nano Banana 2.1 frame plus 6 s Omni 720p is $5.95

A product reel that starts from a generated still and becomes a 6 second clip costs $0.85 at 720p on Sume: $0.10 for a 1K Nano Banana 2.1 image plus 6 x $0.125 = $0.75 for Gemini Omni Flash 1.1. Seven reels are 7 x $0.85 = $5.95. At 1080p the clip is 6 x $0.1875 = $1.125, so a reel is $1.225 and seven are $8.575, which bills as about $8.58 after rounding up to the cent.

This post prices the two steps. It does not claim a quality result; look at one reel before you commit to seven.

The price by resolution

Omni Flash 1.1 on Sume takes 3 to 10 seconds, 16:9 or 9:16, with native audio always on, and bills per output second. The image step is the same $0.10 in every row.

Where the money goes

The clip is 88 percent of a 720p reel ($0.75 of $0.85) and 92 percent of a 1080p reel ($1.125 of $1.225). A cheaper still barely moves the total: swapping the $0.10 frame for a $0.025 Grok image saves $0.075 a reel, or $0.525 over seven. Choosing 360p over 720p saves $0.525 a reel.

That ordering suggests a workflow: choose the frame carefully at 1K, run one 360p clip to check motion at $0.325 all-in, and only then run the 720p or 1080p version.

• Frame: 1 x $0.10 = $0.10 (Nano Banana 2.1, 1K)
• Clip: 6 s x $0.125 = $0.75 (Omni 720p)
• Seven reels: $5.95 at 720p, $8.575 at 1080p, $2.275 at 360p

Before you run it

Read the model descriptors in GET /v1/videos/models for how Omni accepts an image, because fields differ by model, and set 9:16 or 16:9 yourself since Omni supports only those two. Video jobs are asynchronous, so poll the job rather than waiting on the request. Failed generations are not billed, so a retry after a failure does not double the line.

A staged plan for seven reels

Stage one is seven frames at $0.10, which is $0.70. Pick the best frames before spending on video. Stage two is a single 6 second clip at 360p on the best frame, $0.225, to check motion. Stage three is seven clips at 720p: 7 x $0.75 = $5.25. The total is $0.70 + $0.225 + $5.25 = $6.175, about 23 cents more than the $5.95 straight-through price, in exchange for one cheap look before the big spend.

If one frame fails the motion check, swap it and pay only for its clip again. A failed generation is not billed, but a clip that completes and looks wrong is, so the 360p test is the cheaper place to find the problem.

What is not in the number

The price above covers the image and the clip. It does not include music, narration or captions. Omni Flash 1.1 on Sume has native audio always on, so the clip carries its own sound. If you add a voiceover, Sume's TTS is $0.0475 per 1,000 characters, so a 300-character line adds 0.3 x $0.0475 = $0.01425 per reel, or about $0.10 over seven.