Two things this family does that nothing else here does: it generates video from an audio track, and it puts a real draft→final ladder inside one model — the same prompt runs on fast and pro, so promoting a take costs a re-render, not a re-write.
Same prompt, same seed of an idea, one tier apart. Both rendered for this page.
fast · ltx-25-t2v-fast · 6s @ 720p · rendered in 28.1s · $0.57
pro · ltx-25-t2v-pro · 6s @ 720p · rendered in 32.4s · $0.76
ltx-25-a2v-pro — the audio came first; the motion is generated to it, not synced to it afterwards. Billed per second of input audio.
All six endpoints baselined on the production MCP before shipping — 16/16.
| Capability | Probe | Result | Wall clock |
|---|---|---|---|
ltx-25-t2v-fast | 6s · 720p · n=3 | 3/3 | p50 28.1s · p95 30.3s |
ltx-25-t2v-pro | 6s · 720p · n=3 | 3/3 | p50 32.4s · p95 32.6s |
ltx-25-i2v-fast | 6s · 720p · n=3 | 3/3 | p50 33.0s · p95 42.9s |
ltx-25-i2v-pro | 6s · 720p · n=2 | 2/2 | p50 33.1s · p95 35.9s |
ltx-25-a2v-fast | ~6s voice · n=3 | 3/3 | p50 32.7s · p95 39.4s |
ltx-25-a2v-pro | ~6s voice · n=2 | 2/2 | p50 39.7s · p95 41.0s |
| Duration barely moves it: the same fast t2v at 20s took 42.9s — ~22s fixed + ~1s per second of video. | |||
For comparison, Seedance 2.5 needs ~190s for a 4-second take. At half a minute a render, showing someone three variations here costs less wall-clock than one draft there — which is what makes the ladder usable rather than theoretical.
fast reaches 1440p, 4K and 20s takes; pro stops at 1080p and 10s. Need 4K or a long take and fast is the only option.