Reference
Models
Every model and endpoint ID available on the Cheaper Gen. API — video and image — with per-unit prices and links to each model's API docs.
Every model exposes one endpoint per task. Submit to POST /api/v1/run/{endpoint} with the endpoint ID below — the request shape is the same for every model, only the input fields differ. Each model's API page lists its fields, types, defaults and allowed values.
This list is generated from the same catalog the API uses. To fetch it programmatically, call GET /api/v1/models (no key needed) — see List models.
Video models#
Priced per second of output video. Prices shown are the lowest tier; see pricing for every resolution.
Seedance 2.0#
ByteDance Seedance 2.0 generates cinematic 4–15s videos with synchronized audio from text, images or reference assets, at 480p up to 4K.
Reference to Video takes up to 9 images, 3 videos (15 s in total) and 3 audio clips — mention them as @Image1, @Video1, @Audio1. See reference inputs.
Seedance 2.5#
Seedance 2.5 is ByteDance's latest video model with stronger motion, prompt adherence and multi-reference control, generating up to 1080p clips with audio.
Reference to Video takes up to 30 images, 10 videos (30 s in total) and 10 audio clips — mention them as @Image1, @Video1, @Audio1. See reference inputs.
Seedance 2.0 Fast#
Seedance 2.0 Fast is a speed-optimized Seedance 2.0 for 480p and 720p videos with audio — ideal for iteration and high-volume pipelines.
Reference to Video takes up to 9 images, 3 videos (15 s in total) and 3 audio clips — mention them as @Image1, @Video1, @Audio1. See reference inputs.
Seedance 2.0 Mini#
Seedance 2.0 Mini generates 480p and 720p videos with audio at the lowest price in the Seedance family — ideal for previews, A/B tests and volume.
Reference to Video takes up to 9 images, 3 videos (15 s in total) and 3 audio clips — mention them as @Image1, @Video1, @Audio1. See reference inputs.
Wan 3.0#
Alibaba Wan 3.0 generates up to 30 seconds of continuous 480p–1080p video with audio from text, an image or reference images.
Reference to Video takes up to 4 images, 5 videos (15 s in total) and 5 audio clips — mention them as @Image1, @Video1, @Audio1. See reference inputs.
Wan 3.0 Prime#
Wan 3.0 Prime is Alibaba's premium Wan 3.0 tier with higher visual fidelity and motion quality, up to 30 seconds at 1080p with audio.
Reference to Video takes up to 4 images, 5 videos (15 s in total) and 5 audio clips — mention them as @Image1, @Video1, @Audio1. See reference inputs.
MiniMax H3#
MiniMax H3 (Hailuo 3) generates expressive 5–15s videos with strong physics and character motion at 768P or 2K, from text, images or references.
Reference to Video takes up to 5 images, 3 videos (15 s in total) and 3 audio clips — mention them as @Image1, @Video1, @Audio1. See reference inputs.
MiniMax H3 Max#
MiniMax H3 Max is the post-trained, higher-quality H3 with sharper detail and stronger prompt adherence, generating 480P and 768P videos of 5–15 seconds.
Image models#
Priced per generated image. Prices shown are the lowest tier; see pricing for every resolution.
Seedream 5.0 Pro#
ByteDance Seedream 5.0 Pro creates and edits photorealistic, on-brand images up to 2K with precise multi-image editing and accurate text rendering.
Seedream 5.0 Lite#
Seedream 5.0 Lite delivers strong Seedream 5.0 image quality and editing at a lower price — great for everyday creative and app workloads.
Seedream 5.0 Flash#
Seedream 5.0 Flash is the fastest and most affordable Seedream 5.0 model for real-time and high-volume image generation and editing.
Text models#
LLM inference — DeepSeek, Qwen, GLM, Kimi and more — is coming to the same API and balance. Join the waitlist to hear when it launches.
Something unclear or missing? Email support@cheapergen.com — we read everything.