60-second AI short film, narrated

The whole thing: script, keyframes, clips and a voiceover. Every stage a measured first-party number, nothing assumed.

The pipeline

  1. Write the script and shot list, qwen3-32b
  2. Generate 15 keyframes, stable-diffusion-xl
  3. Render 15 clips (~4s each), ltx-video
  4. Narrate 60 seconds of voiceover, kokoro-tts

Fastest card where every stage is a real measurement: NVIDIA H100 80GB HBM3, 2.6 min for the whole job.

Every GPU, slowest job to fastest

GPUTotal timeComputeEnergyBasis
NVIDIA H100 80GB HBM32.6 min2 min17.46 Whall 4 stages measured
NVIDIA H2004.1 min2 min18.04 Whall 4 stages measured
NVIDIA A100 40GB SXM44.2 min3.9 min6.4 Whanchored estimate (3/4 stages measured)
NVIDIA A100 80GB SXM44.9 min4 min23.62 Whall 4 stages measured
NVIDIA L40S5.1 min4.3 min22.53 Whall 4 stages measured

Cards that can't run this pipeline

Every one of these fails on the same kind of wall, a stage that will not fit in VRAM.

…and 18 more.

How these numbers are built

Each stage time is the quantity of work divided by that card's measured throughput for that model, from our own bench. The pipeline is assumed to run batched, every image, then every clip, so each model loads once. Model load time is added where we recorded it; our text-generation runs don't carry a load measurement yet, so pipelines with a language-model stage are slightly optimistic. Nothing here is a single timed run of the whole pipeline, and we don't present it as one.