Battle · Image-to-Video · voting is not open yet

Runway Gen-4.5 vs Ray3.2

No votes yet — voting is not open yet. This page compares what the two vendors publish about their own models; it records no winner, no win share and no score.

Open the AIContentDrop generator →
Runway Gen-4.5 Proprietary
VendorRunway [1]
LicenceProprietary [1]
Release2025-12-01 [1]
Max resolutionNot stated [1]
Max clip durationNot stated [1]
Native audioNot stated [1]

Runway Gen-4.5 model page →

Ray3.2 Proprietary
VendorLuma AI [2]
LicenceProprietary [2]
Release2026-06-09 [2]
Max resolution1080p [2]
Max clip duration20 s [2]
Native audioNot stated [2]

Ray3.2 model page →

Side-by-side published facts

Runway Gen-4.5 and Ray3.2 — vendor-sourced, checked 2026-10-06
FactRunway Gen-4.5 Ray3.2
VendorRunway [1]Luma AI [2]
LicenceProprietary [1]Proprietary [2]
Release2025-12-01 [1]2026-06-09 [2]
ModesText-to-video, Image-to-video, Video edit [1]Text-to-video, Image-to-video, Video edit [2]
Max resolutionNot stated [1]1080p [2]
Max clip durationNot stated [1]20 s [2]
Native audioNot stated [1]Not stated [2]
AccessAPI, web app, mobile app [1]API, web app [2]
Arena ratingNo votes yetNo votes yet

What the table does not settle

Two hosted creative platforms with the thinnest published specifications in the catalogue. Luma states 1080p and twenty seconds for Ray3.2 but says nothing we could read about native audio [2]; Runway states neither a resolution nor a duration for Gen-4.5 [1]. Four of the sixteen value cells in this comparison are Not stated — Gen-4.5's maximum resolution, maximum clip duration and native audio, plus Ray3.2's native audio — and we would rather show that than fill them in.

What both vendors do describe is a posture: Luma pitches frame-by-frame direction, Runway pitches a model available on every paid plan and cites a top placement on a third-party benchmark. Neither claim is a measurement this site has made, and neither will be until voters have compared the clips themselves.

Run the same brief through both models yourself before anyone's leaderboard tells you which to use: try both models in one workspace · estimate credits for the pair.

Sources

  1. Runway — Runway Gen-4.5, vendor documentation https://runway.com/research/introducing-runway-gen-4.5 — checked 2026-10-06
  2. Luma AI — Ray3.2, vendor documentation https://lumalabs.ai/news/introducing-ray-3-2 — checked 2026-10-06

Each [n] above is the vendor page the fact beside the marker was read from, on the date shown. Where a vendor states nothing, the table says “Not stated” and no value is inferred from anywhere else.