LTX-2.3 4K at $0.24/s on fal vs Omni 4K at $0.375/s on Sume

A 10-second 4K clip costs $2.40 with LTX-2.3 on fal ($0.24 a second) and $3.75 with Gemini Omni Flash 1.1 on Sume ($0.375 a second). Sume does not list an LTX id, so if you need 4K through the Sume API, Omni is the 4K id in the catalog.

The two prices are not the same kind of number. The fal figure is a provider list price. The Sume figure is the provider list price times 1.25, the house margin that applies to every Video Router row.

The arithmetic

fal's LTX-2.3 page lists text-to-video and image-to-video at $0.06 a second at 1080p, $0.12 at 1440p and $0.24 at 2160p, with cheaper Fast variants and a 20-second maximum.

Sume's catalog lists Omni at $0.03, $0.10, $0.15 and $0.30 a second at 360p, 720p, 1080p and 4K. Multiply each by 1.25 and you get $0.0375, $0.125, $0.1875 and $0.375. So 4K is 10 x $0.375 = $3.75 for the longest Omni clip, which is 10 seconds.

What the price does not tell you

LTX-2.3 gives you 20 seconds in one generation. Omni stops at 10 seconds, so a 20-second 4K piece on Sume is two Omni jobs, which is $7.50 against $4.80, plus a join. If length is the point, that gap matters more than the per-second rate.

Omni also fixes two things you may care about. Native synced audio is always on, and generate_audio: false is rejected. The aspect ratio is 16:9 or 9:16 only. See the 4K model list by resolution before you commit a budget.

Request a 4K Omni clip on Sume

Omni accepts 4K in resolution, and lowercase 4k as an alias on this model only. The default is 720p, so you must ask for 4K.

curl -X POST https://api.sume.com/v1/videos \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: omni-4k-001" \
  -d '{"model":"gemini-omni-flash-1.1","prompt":"Slow dolly across a ceramic mug on an oak desk, morning light","resolution":"4K","duration":10,"aspect_ratio":"16:9"}'

When to pick which

Pick LTX-2.3 when you want the cheapest per-second 4K, 20-second takes, or open weights you can run yourself. Running weights yourself means you pay for hardware instead of per second.

Pick Omni on Sume when you want one hosted API for 4K, edits and references, a single bill, and job polling that is the same as every other Sume video model. Read more about the Omni modes in the Omni API walkthrough.

Checklist before you commit a 4K budget

Price per second is only one input. Before you plan a batch, run through these points.

• Length: LTX-2.3 gives 20 seconds per generation on fal, Omni gives 10 on Sume, so count jobs, not just seconds.
• Audio: both produce native audio, but Omni's cannot be turned off on Sume.
• Shape: Omni takes 16:9 and 9:16 only. If you need 1:1 or 4:3 at 4K, neither route in this post is a fit.
• Delivery: check the file at full size on the screen where it will play. A 4K label does not guarantee more visible detail than a clean 1080p clip.
• Spend: send one short 4K test first, read usage.cost on the finished job, and compare it with the table before you queue the batch.