Video Models
Video models in Armox transform text descriptions and images into dynamic video content. From short clips to cinematic productions, these models bring your ideas to life.
Overview
Video models can:
- Text-to-video — Generate videos from text descriptions
- Image-to-video — Animate still images
- Motion control — Apply specific movements and camera work
- Audio generation — Some models include native audio
- Multi-shot editing — Create connected video sequences
Available Video Models
| Model | Provider | Cost | Duration | Best For |
|---|---|---|---|---|
| Kling V2.5 Turbo Pro | Kuaishou | 1,000 credits | 5-10s | Cinematic quality |
| Kling 2.6 Pro | Kuaishou | 1,200 credits | 5-10s | Native audio |
| Veo 3.1 | 4,000 credits | 4-8s | Premium quality | |
| Veo 3.1 Fast | 1,500 credits | 4-8s | Fast generation | |
| Seedance 1 Pro | ByteDance | 1,000 credits | 2-12s | Long videos |
| Seedance 1 Pro Fast | ByteDance | 500 credits | 2-12s | Budget option |
| Seedance 1 Lite | ByteDance | 300 credits | 2-10s | Dance/motion |
| Wan 2.6 I2V | Wan | 1,000 credits | 5-15s | Audio sync |
| Wan 2.6 T2V | Wan | 1,000 credits | 5-15s | Text-to-video |
| Sora 2 Pro | OpenAI | 4,000 credits | 4-12s | Premium quality |
| Sora 2 | OpenAI | 2,000 credits | 4-12s | Good quality |
| Hailuo 2.3 | MiniMax | 1,120 credits | 6-10s | Prompt optimization |
| Hailuo 2.3 Fast | MiniMax | 600 credits | 6-10s | Fast generation |
| Video-01 Director | MiniMax | 1,000 credits | Variable | Camera control |
| Gen4 Aleph | Runway | 1,000 credits | 5s | Video editing |
Connection Colors
In the Armox Canvas, video connections use green handles and edges:
- Input Handle: Green circle on the left side of nodes
- Output Handle: Green circle on the right side of nodes
- Connection Edge: Green line connecting nodes
Common Settings
Aspect Ratio
- 16:9 — Landscape (YouTube, presentations)
- 9:16 — Portrait (TikTok, Reels, Stories)
- 1:1 — Square (Instagram feed)
Duration
Most models support 5-10 second clips. Some offer longer durations.
Resolution
- 720p — Standard quality, faster
- 1080p — High quality, recommended
- 4K — Maximum quality (some models)
Choosing the Right Model
For Maximum Quality
- Veo 3.1 (4,000 credits) — Google's best
- Sora 2 Pro (4,000 credits) — OpenAI's premium
- Kling 2.6 Pro (1,200 credits) — Excellent with audio
For Speed and Value
- Seedance 1 Pro Fast (500 credits) — Good balance
- Hailuo 2.3 Fast (600 credits) — Quick results
- Seedance 1 Lite (300 credits) — Budget option
For Audio
- Kling 2.6 Pro — Native audio generation
- Veo 3.1 — Audio support
- Wan 2.6 models — Audio synchronization
For Image Animation
- Wan 2.6 I2V — Image-to-video specialist
- Kling models — Strong image animation
- Seedance models — Good motion consistency
Best Practices
- Start with shorter durations — Test concepts first
- Use negative prompts — Avoid unwanted elements
- Describe motion — Be specific about movement
- Consider audio needs — Choose models with audio if needed
- Match aspect ratio to platform — 9:16 for social, 16:9 for YouTube
Next Steps
Explore individual model documentation for detailed settings and use cases.