# Model Selection and Configuration Qolaba provides access to multiple state-of-the-art video generation models, each with different strengths across realism, speed, creative quality, and supported capabilities. Choosing the right model before generating directly affects output quality, available controls, and credit usage. *** #### How to Select a Model 1. Open the **Video Generation** workspace 2. Click the **Model Selector** at the top of the interface 3. Select your preferred model — the credit cost per generation updates in the bottom bar immediately *** #### Available Models | Model | Strengths | Max Duration | Max Quality | Audio Control | | ---------------- | -------------------------------------------------------------------------------------- | ------------ | ----------- | ------------- | | **Veo 3.1** | Google's flagship model — high realism, physics-accurate motion, detailed environments | 8s | 4K | — | | **Veo 3.1 Fast** | Faster generation of Veo 3.1 — balanced speed and quality | 8s | 4K | — | | **Seedance** | Strong reference-guided generation — supports image, video, and audio references | 8s | 4K | ✓ | | **Seedance 2.5** | Updated Seedance with improved motion consistency and reference handling | 8s | 4K | ✓ | | **Happy Horse** | Creative and stylized outputs — supports reference image input | 8s | 1080p | — | | **Runway** | High production quality — reliable for commercial and branded content | 8s | 1080p | — | | **Hailuo** | Smooth motion, strong for cinematic storytelling | 6s | 1080p | — | | **Kling 3** | Fast short-form video — cost-efficient, good for social media clips | 8s | 4K | — | | **Kling V3** | Updated Kling with improved motion realism and detail | 8s | 4K | — | | **Vidu Q3** | Supports very short duration clips — ideal for quick cuts and transitions | 1s+ | 1080p | — | | **Luma** | Cinematic quality, smooth motion — strong for creative and narrative content | 8s | 4K | — | | **Imagine** | Creative and experimental outputs — good for stylized and artistic generation | 8s | 1080p | — | *** #### Model-Specific Capabilities Not all models support the same input and output features. The table below shows what each model supports for reference media uploads: | Model | Initial Image | Last Frame Image | Reference Files (Image, Video, Audio) | | ---------------- | ------------- | ---------------- | ------------------------------------- | | **Veo 3.1** | ✓ | ✓ | — | | **Veo 3.1 Fast** | ✓ | ✓ | — | | **Seedance** | ✓ | ✓ | ✓ Image, Video, Audio | | **Seedance 2.5** | ✓ | ✓ | ✓ Image, Video, Audio | | **Happy Horse** | ✓ | — | ✓ Image only | | **Runway** | ✓ | — | — | | **Hailuo** | ✓ | ✓ | — | | **Kling 3** | ✓ | ✓ | — | | **Kling V3** | ✓ | ✓ | — | | **Vidu Q3** | ✓ | — | — | | **Luma** | ✓ | ✓ | — | | **Imagine** | ✓ | — | — | {% hint style="info" %} All models support uploading an **initial image** as the starting frame. Last frame and reference file support varies — check the table above before selecting a model if these inputs are part of your workflow. {% endhint %} *** #### Model Recommendations by Use Case 1. **Highest realism and motion accuracy** → Veo 3.1 — Google's most advanced model with physics-accurate motion and highly detailed environments. Best for cinematic and documentary-style content. 2. **Fastest generation** → Veo 3.1 Fast or Kling 3 — Balanced speed and quality for workflows where turnaround time matters more than maximum fidelity. 3. **Reference-guided generation** → Seedance or Seedance 2.5 — The only models that support image, video, and audio as reference inputs. Best for generation that needs to stay consistent with existing creative assets. 4. **Cinematic storytelling** → Luma or Hailuo — Smooth, cinematic motion with strong temporal consistency. Well-suited for narrative content, mood reels, and creative storytelling. 5. **Short-form social content** → Kling 3 or Kling V3 — Cost-efficient, fast, and well-optimized for short clips with strong motion. 6. **Stylized and artistic output** → Happy Horse or Imagine — Creative and experimental results with a more stylized aesthetic. Good for branded or artistic content that doesn't require photorealism. 7. **Commercial and branded production** → Runway — Reliable, high-quality output suited for commercial use cases with consistent results. 8. **Ultra-short clips and transitions** → Vidu Q3 — Supports durations starting from 1 second. Ideal for quick cuts, looping visuals, and short transitions. 9. **Audio-included video generation** → Seedance or Seedance 2.5 — The only models with built-in audio generation controls. All other models generate video without audio. *** #### Credit Cost Visibility Each model displays its credit cost per generation next to its name in the model selector. Credit cost varies based on: * The model selected * Resolution and quality setting * Video duration * Audio on or off (Seedance models only) * Number of generations The **live credit counter** in the bottom bar updates in real time as you adjust any of these settings, so you always know the exact cost before clicking Generate. For a full breakdown of how credits are calculated, see [Understanding Credit Usage →](/video-generation/understanding-credit-usage.md) --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.qolaba.ai/video-generation/model-selection-and-configuration.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.