Lightricks models, access and benchmark guide
Lightricks model and access guide with a practical US-English evaluation brief and links to publisher materials.

👁 111 reads
A creative team might find LTX through Lightricks’ production interface or the public LTX-Video repository. The practical difference is who supplies compute and maintains the runtime. I would name the execution environment and checkpoint before estimating a local project.
Choosing the first job
Film a measuring tape retracting and inspect whether the metal hook remains connected to the tape.
Models in the November 17 snapshot
| Model and track | Placement | Reported score |
|---|---|---|
| LTX-2.5 Fast (AA text to video) | 29 | 939 Elo |
| LTX-2.5 Fast (AA image to video) | 32 | 912 Elo |
| LTX-2.5 Pro (AA text to video) | 30 | 938 Elo |
| LTX-2.5 Pro (AA image to video) | 31 | 912 Elo |
| LTX-2.3 Pro (AA text to video) | 33 | 885 Elo |
| LTX-2.3 Pro (AA image to video) | 34 | 837 Elo |
| LTX-2.3 Fast (AA text to video) | 34 | 836 Elo |
| LTX-2.3 Fast (AA image to video) | 35 | 832 Elo |
Release record. Publisher documentation identifies the product family. The ranked configurations include LTX-2.5 Fast, LTX-2.5 Pro, LTX-2.3 Pro.
Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →Sources
- https://ltx.io/ (accessed November 17, 2026)
- https://github.com/Lightricks/LTX-Video (accessed November 17, 2026)
- https://artificialanalysis.ai/video/leaderboard/text-to-video (accessed November 17, 2026)
- https://artificialanalysis.ai/video/leaderboard/image-to-video (accessed November 17, 2026)
More on this from QUANTSLANT
ai tools·Language models ranked by intelligence and task costRead the Intelligence Index next to evaluation cost, then choose a reasoning setting that fits the US workflow you actually run.