What are Models
Understanding AI models on WaveSpeedAI.
Overview
Models are AI systems trained to generate specific types of content. WaveSpeedAI hosts 1,000+ models from various providers, all accessible through a unified API.
Model Categories
| Category | Description | Example Models |
|---|---|---|
| Text-to-Image (T2I) | Generate images from text | Z-Image Turbo, Seedream, Nano Banana |
| Image-to-Image (I2I) | Transform or edit images | FLUX Kontext, Seedream Edit |
| Text-to-Video (T2V) | Create videos from text | Wan T2V, Kling T2V |
| Image-to-Video (I2V) | Animate images | Veo 3.1, Kling I2V, Hailuo |
| Video-to-Video (V2V) | Transform or extend videos | Video effects, style transfer |
| Text-to-Audio | Generate speech, music, sounds | Minimax Speech, ElevenLabs, Dia TTS |
| Audio-to-Video | Generate video from audio | MMAudio |
| Image-to-Text | Describe or analyze images | Molmo, Qwen VL |
| Image-to-3D | Create 3D models from images | Hunyuan3D, Tripo3D |
| Training | Train custom LoRA models | FLUX LoRA Training |
Browse all categories at wavespeed.ai/models.
Model Identifiers
Models exposed through the generation API use the complete model ID shown on the model page. Most current generation models use three segments:
{vendor}/{model}/{variant}Examples:
wavespeed-ai/z-image/turbopixverse/pixverse-v5.5/image-to-videowavespeed-ai/hunyuan-video/t2v
Some models use two segments, such as minimax/speech-2.6-hd. Always use the exact ID shown on the model page rather than adding or removing a segment. Model IDs used by the OpenAI-compatible LLM API are a separate namespace and may use a different format.
Model Properties
Parameters
Each model has its own set of input parameters. Here are some common examples:
| Parameter | Description | Used By |
|---|---|---|
prompt | Text description of what to generate | Most models |
image_url | Input image URL | I2V, I2I models |
width, height | Output dimensions | Image models |
duration | Video length in seconds | Video models |
seed | Random seed for reproducibility | Most models |
Important: Always check the model’s README for its specific parameters. Each model page shows all available parameters with descriptions and defaults.
Pricing
Models have different pricing based on:
- Model complexity
- Output type and size
- Provider costs
Rate Limits
Some models may have additional rate limits beyond your account tier.
Finding Models
Browse
Visit wavespeed.ai/models to browse all models with filters.
API
List available models programmatically:
curl --fail-with-body --connect-timeout 10 --max-time 60 \
"https://api.wavespeed.ai/api/v3/models" \
-H "Authorization: Bearer $WAVESPEED_API_KEY"Model Documentation
Each model has detailed documentation including:
- Parameters and defaults
- Example requests
- Pricing information
- Best practices
See the Model Library for complete documentation.