Skip to content

Which AI Video Model Is Cheapest? I Tested 18

8 min

After I published my previous AI image generation experiment, I thought, why not benchmark text-to-video models available through Vercel AI Gateway as well.

This time I added some action to the puppy prompt:

A golden retriever puppy chasing a butterfly through a sunflower field at golden hour, cinematic shallow depth of field, soft warm light, slow camera dolly forward.

Tried to make the models consistently produce clips of 1280x720 (720p), 16:9, 4-5s, with audio where possible.

The results below are from runs on May 17-18, 2026. Prices and model availability may have changed since.

Jump straight to the winners if you don't have much time, or check the data below.

Full Results

Model Cost Generation time Date Duration Resolution Audio Video
alibaba/wan-v2.5-t2v-preview $0.5 115.3s 2026-05-18 5.0s 1280x720 yes video
alibaba/wan-v2.6-t2v $0.5 53.4s 2026-05-17 5.0s 1280x720 yes video
bytedance/seedance-v1.0-pro $0.2575 53.0s 2026-05-18 5.0s 1248x704 no video
bytedance/seedance-v1.0-pro-fast $0.103 44.3s 2026-05-18 5.0s 1248x704 no video
bytedance/seedance-v1.5-pro $0.1295 73.9s 2026-05-18 5.0s 1280x720 yes video
bytedance/seedance-2.0 $0.7623 170.4s 2026-05-17 5.1s 1280x720 yes video
bytedance/seedance-2.0-fast $0.60984 117.9s 2026-05-18 5.1s 1280x720 yes video
google/veo-3.0-generate-001 $1.6 85.1s 2026-05-18 4.0s 1280x720 yes video
google/veo-3.0-fast-generate-001 $0.6 64.1s 2026-05-18 4.0s 1280x720 yes video
google/veo-3.1-generate-001 $1.6 54.6s 2026-05-17 4.0s 1280x720 yes video
google/veo-3.1-fast-generate-001 $0.6 65.6s 2026-05-18 4.0s 1280x720 yes video
klingai/kling-v2.5-turbo-t2v $0.211722 47.8s 2026-05-18 5.0s 1280x720 no video
klingai/kling-v2.5-turbo-t2v-pro $0.35287 81.5s 2026-05-18 5.0s 1920x1080 no video
klingai/kling-v2.6-t2v $0.211722 57.0s 2026-05-18 5.0s 1280x720 no video
klingai/kling-v2.6-t2v-pro $0.70574 90.9s 2026-05-18 5.0s 1920x1080 yes video
klingai/kling-v3.0-t2v $1.270332 41.6s 2026-05-17 5.0s 1280x720 yes video
klingai/kling-v3.0-t2v-pro $1.693776 83.0s 2026-05-18 5.0s 1920x1080 yes video
xai/grok-imagine-video $0.35 49.6s 2026-05-17 5.0s 1280x720 yes video

Total spent: $12.058302

  • Each model generated one video, so each generation time is a single measurement.
  • Generation time is wall time per video, measured by the benchmark script.
  • Cost is returned by the gateway, so it should be accurate.
  • Date is when each row's video was produced - rows may be from different runs (cache hits aren't re-generated).
  • Duration, Resolution and Audio are measured from each saved video, not taken from the requested settings. Costs are only like-for-like between rows where these match.

Cost

AI video generation cost chart

Generation Time

AI video generation latency chart

Model Quirks

A couple of observations from running this benchmark:

  • The -fast variants from ByteDance and Google trade a bit of quality for significantly lower cost - Veo 3.x Fast is ~62% cheaper than the standard Veo 3.x at $0.6 vs $1.6.
  • Kling Pro variants are consistently slower and pricier than their base counterparts. Part of the price is resolution: the Pro variants rendered 1920x1080, while the base ones rendered 1280x720. I didn't see an obvious quality jump in this scene.
  • Not every model returned what I requested: Seedance v1.0 Pro and Pro Fast rendered 1248x704, and Veo returned 4-second clips instead of 5.
  • Some models returned no audio: Seedance v1.0 Pro and Pro Fast, Kling 2.5 Turbo (both variants), and Kling 2.6 (base).

Winners

If you only optimize for cost, the cheapest models in this benchmark are:

  • bytedance/seedance-v1.0-pro-fast at $0.103 (1248x704, no audio)
  • bytedance/seedance-v1.5-pro at $0.1295 (the cheapest with 720p and audio)
  • klingai/kling-v2.5-turbo-t2v and klingai/kling-v2.6-t2v tied at $0.211722 (both without audio)

If you optimize for latency, the fastest models are:

  • klingai/kling-v3.0-t2v at 41.6s
  • bytedance/seedance-v1.0-pro-fast at 44.3s
  • klingai/kling-v2.5-turbo-t2v at 47.8s

But to be honest, when we're talking about video, the scene, prompt-following precision, and quality matter more than cost and latency, so I'll let you choose the winner yourself.

DIY

If you want to run the benchmark yourself, add more models, or adjust the prompt - the script is easy to configure and run: kometolabs/ai-video-generation-cost-analysis

Support My Work

Preparing such research sometimes takes 2-5 full benchmark runs, which adds up. If it saved you time, consider sponsoring me on GitHub or Buy Me a Coffee.

Relevant Posts