AI Models Guide

vpick integrates multiple AI models covering image, video, voice, and music generation. Here's a detailed guide and comparison.

Image Models

Nano Banana 2 (Default)

Nano Banana 2 is vpick's default image model, offering a great balance of quality, speed, and price. Supports multiple reference image inputs, ideal for product photos, creative designs, and more.

Supported aspect ratios: 1:1, 16:9, 9:16, 4:3, 3:4

Grok Imagine

Grok Imagine is budget-friendly — text-to-image produces 6 images at once, and image-to-image (requires connected reference images) produces 2 images. Perfect for creative brainstorming with lots of variations.

Seedream

Seedream is the cheapest image model available, ideal for bulk generation or budget-conscious projects.

Supported aspect ratios: 1:1, 16:9, 9:16, 4:3, 3:4, 21:9

Two resolution options:

• Seedream (2K): Standard resolution
• Seedream HD (3K): High resolution

Video Models

Veo 3.1 Fast (Default)

Veo 3.1 Fast is vpick's default video model with exceptional visual quality and built-in sound generation. Ideal for high-quality showcase content.

Kling 3.0

Kling 3.0 is reliable with flexible duration from 3 to 15 seconds, plus start/end frame control and sound generation. Pro mode outputs 1080p.

Supported aspect ratios: 1:1, 16:9, 9:16

Advanced: MultiShot

MultiShot mode lets you split one video into 1-5 segments, each with an independent prompt and duration (1-12 seconds), totaling 3-15 seconds. Sound is automatically enabled when MultiShot is active. Requires a connected start frame image.

Advanced: Elements

Elements lets you define character or object references with images, then reference them in prompts using @element_name. Each element can have 2-50 reference images, and the AI will maintain consistent character appearance. If no start frame image is connected, the first element's images are used automatically.

Grok Video

Grok Video offers flexible duration options (6/10/15 seconds), 480p and 720p quality, and sound generation. Reasonably priced, great for social media content.

Generation modes: Fun (creative), Normal (balanced), Spicy (dynamic)

Runway Gen4

Runway Gen4 is known for outstanding video quality and is one of the few models supporting 1080p. 5-second clips are the most reliable.

Supported aspect ratios: 16:9, 4:3, 1:1, 3:4, 9:16

Video Model Comparison

Recommendations:

• Best quality: Veo 3.1 Fast (default)
• Need sound: Veo 3.1 Fast, Kling 3.0, or Grok Video
• Best value: Grok Video (480p-6s $0.15) or Runway Gen4 (720p-5s $0.18)
• Longer videos: Kling 3.0 (up to 15s) or Grok Video (up to 15s)
• High resolution: Kling 3.0 Pro or Runway Gen4 (1080p)

Voice Model

ElevenLabs Text-to-Dialogue V3

ElevenLabs V3 is one of the most natural AI voice synthesis models available.

Available voices:

Supported languages: Auto-detect, English, Chinese, Japanese, Korean, Spanish, French, German, Portuguese, Italian

Parameters:

• Stability: 0.0-1.0 — higher values produce more consistent voice, lower values add more expressiveness and emotion

Music Model

Suno V4.5

Suno V4.5 generates complete music tracks from text descriptions with professional-grade quality.

Modes:

• Simple mode: Enter a text description, Suno automatically decides style and lyrics
• Custom mode: Specify music style and song title

Options:

• Instrumental: Enable to generate vocal-free music, perfect for background music
• Vocal Gender: Auto / Male / Female

Lipsync Model

Kling Avatar

Kling Avatar combines a portrait photo and an audio clip into a "talking video" where the person's lip movements precisely match the audio.

How to use:

1. Connect a portrait photo (image-in)
2. Connect an audio clip (audio-in)
3. Run lipsync generation

Best practices:

• Use a front-facing, clear portrait photo
• The person's mouth should preferably be closed
• Better audio quality = better lip sync results

Pricing Overview

Images

Videos

Audio