Model Guide
Video Models
Model Name | Use Case / Notes | Resolution | Duration(s) | Credit Cost(s) |
---|---|---|---|---|
Seedance v1-Lite | Smooth motion, cinematic look, 720p output, optimized for speed and efficiency | 720p / 480p | 5s | 22 |
Seedance Pro | Premium motion quality with better detail and lighting. Ideal for ads, story scenes, and promo visuals. | 540p / 720p / 1080p | 5s | 10 / 34 / 76 |
Minimax Hailuo-02 | Realistic physics, natural motion, high visual coherence, strong temporal consistency | 720p | 6s / 10s | 32 / 54 |
Minimax Hailuo-02 Pro | Enhanced motion realism, better physical accuracy, higher fidelity for complex sequences | 1080p | 6s | 58 |
Vidu Q1 | 1080p quality, fluid movement, strong subject consistency, stable compositions | 1080p | 4s | 60 |
Veo 2 | Rich visuals, accurate subject control, camera motion options, fine detail rendering | 720p | 5–8s | 300–480 |
Kling v2.1 Standard | 720p output, stylized visuals, consistent motion, fast generation | 720p | 5s / 10s | 30 / 60 |
Kling v2.1 Pro | 1080p output, improved realism, better lighting and motion physics | 1080p | 5s / 10s | 54 / 108 |
Kling v2.1 Master | Best-in-class fidelity, cinematic camera moves, top-tier consistency and coherence | 720p | 5s / 10s | 168 / 336 |
Luma ray-1–5 | High realism, strong motion control, evolving fidelity across versions | N/A | 5s | 60 |
Luma ray-flash-2 | Faster inference, upgraded dynamics engine, responsive to input adjustments | 1080p | 5s / 9s | 24 / 43 |
Luma ray-2 | Natural-looking motion, clean subject rendering, stable scene transitions | 1080p / 720p / 540p | 5s / 9s | 60 / 108 |
Runway gen4_turbo | Speed-optimized, smoother transitions, better facial stability | 720p | 5s / 10s | 30 / 60 |
Runway gen3a_turbo | Quick generation, basic motion fidelity, good for simple or abstract sequences | N/A | 5s | 30 |
Image Models
Model Name | Use Case / Notes | Resolution(s) Supported | Credit Cost Range |
---|---|---|---|
GPT-Image-01 (Low) | Fast, low-cost draft visuals. Great for quick iterations. | 1024×1024 / 1024×1536 / 1536×1024 | 1–2 |
GPT-Image-01 (Med) | Balanced detail + speed. Ideal for storyboards, concept art. | 1024×1024 / 1024×1536 / 1536×1024 | 5-8 |
GPT-Image-01 (High) | High-fidelity assets. Best for final frame visuals. | 1024×1024 / 1024×1536 / 1536×1024 | 20-30 |
Kontext Pro | Designed for iterative image creation; maintains character and style consistency across edits. | 1024x1024 | 5 |
Kontext Max | Prioritizes maximum performance and speed; enhanced typography capabilities; ideal for high-volume professional workflows. | 1024x1024 | 10 |
Kontext Multi-image | Merges multiple uploaded photos into a single stylistic base; understands shared features like lighting and composition intent. | Grid / batch layout | 10 |
Seedream v3 | High-resolution (up to 2K) image generation; improved aesthetics and structural quality; supports bilingual (Chinese-English) prompts. | 1024x1024 | 4 |
Imagen 4 Fast | Ultra-fast image generation (up to 10x faster than previous versions); maintains high-quality outputs with fine details. | 1024x1024 | 2 |
Imagen 4 | Generates photorealistic images with fine details; enhanced typography capabilities; supports 2K resolution. | 1024x1024 | 5 |
Imagen 4 Ultra | Produces exceptionally high-quality images; includes advanced features like prompt enhancement and style customization; supports multiple languages. | 1024x1024 | 9 |