Side by side
See the difference.
Scores, capabilities and costs for the models you’re considering.
Scroll to compare models →
| 3 models | fal-aiLatentSyncRemove × | alibabaAlibaba: Wan 3.0Remove × | minimaxMiniMax H3 Max Text to VideoRemove × |
|---|---|---|---|
| What’s different | LatentSync is a video-to-video model that generates lip sync animations from audio using advanced algorithms for high-quality synchronization. | Can choose clip duration from the prompt; toggle generated audio, prompt expansion and extra thinking before rendering. | Uses native 480P/768P generation or 1080P latent refinement; supports supplied soundtracks and selectable prompt expansion. |
| Date | Mar 25, 2025Listed | Aug 19, 2026Released | Aug 26, 2026Released |
API prices USD · provider configuration shown below | |||
| Listed rate | See provider quotes | $0.05/ secondDeepInfra · all routes ↗480P · $ cost per second · output second | See provider quotes |
| Provider endpoints | 0 endpoints → | 2 endpoints → | 0 endpoints → |
Benchmarks Each result retains its evaluation configuration | |||
| AA Video + Audio ↗2.1 | Not reported | 1,174.59 ↗Text-to-video · audio evaluated · all categoriesIndependent result95% CI 1,159.59–1,189.59 · n=6,372 | 1,153.51 ↗Text-to-video · audio evaluated · all categoriesIndependent result95% CI 1,138.51–1,168.51 · n=6,144 |
| AA Video Silent ↗2.1 | Not reported | 1,134.11 ↗Text-to-video · silent evaluation · all categoriesIndependent result95% CI 1,116.11–1,152.11 · n=4,042 | 1,116.88 ↗Text-to-video · silent evaluation · all categoriesIndependent result95% CI 1,099.88–1,133.88 · n=5,883 |
Specifications | |||
| Context window | — | — | — |
| Max output | — | — | — |
| Parameters | — | — | — |
| Structured output | Not reported | Supported | Not reported |
| Image input | Not reported | Supported | Not reported |
| Audio input | Not reported | Supported | Not reported |
| Video input | Not reported | Supported | Not reported |