SOpen

LTX-2.3

Lightricks

The open-source flagship. First open model to close the gap with proprietary leaders.

SClosed

Wan 2.6

Alibaba (Tongyi Lab)

Closed-source evolution of Wan. Adds reference-to-video for character consistency, multi-shot narratives, 5 aspect ratios, and 15s duration. Native audio with synced dialogue carried over from 2.5.

Pick LTX-2.3 if…

You want open-source audio-video, or studios needing IP control.

Pick Wan 2.6 if…

You want cross-platform content (all aspect ratios), character-consistent narratives (ref-to-video), or audio-synced social content.

Specifications

Maker
Lightricks
Alibaba (Tongyi Lab)
Source Type
Open Source
Closed Source
License
Apache 2.0 (<$10M rev)
Alibaba Commercial
Architecture
DiT (22B) + Rebuilt VAE
DiT + MoE (evolved)
Parameters
22B
Undisclosed
Max Resolution
4K
720p / 1080p
Max Duration
20s
Up to 15s
FPS
Up to 50
24
Native Audio
Yes
Yes
ComfyUI Support
Yes
Yes
Fine-tunable
Yes
No
Min VRAM
32GB (full) / 12GB (distilled)
Cloud / API
Cost / Second
$0.04
$0.05
Inputs
T2V, I2V, V2V, Audio-cond
T2V, I2V, Reference-to-Video (1-3 refs via @Video1/@Video2/@Video3)
On Floyo
Yes
Yes

Strengths & Trade-offs

LTX-2.3

Strengths

  • +22B params
  • +true 4K at 50fps
  • +first open model with synced audio
  • +rebuilt VAE
  • +native portrait

Trade-offs

  • -Full 4K needs 48GB
  • -dialogue lip-sync inconsistent
  • -in-scene text flaky

Best For

  • Local 4K production
  • open-source audio-video
  • studios needing IP control

Wan 2.6

Strengths

  • +Fastest inference
  • +native audio with synced dialogue
  • +reference-to-video for character consistency (1-3 video refs)
  • +multi-shot with structured prompt syntax [0-3s]/[3-5s]
  • +expanded aspect ratios (16:9, 9:16, 1:1, 4:3, 3:4)

Trade-offs

  • -Closed source (not self-hostable)
  • -reference-to-video limited to 5/10s (no 15s)
  • -800 char prompt limit
  • -multi-shot timing depends on prompt expansion quality
  • -check regional license terms

Best For

  • Cross-platform content (all aspect ratios)
  • character-consistent narratives (ref-to-video)
  • audio-synced social content
  • multilingual production