The model library
AI video models, all in one place
The new Seedance 2.5, plus MiniMax H3, Kling, Wan and Happy Horse — the leading AI video models, online on DreamFort with one prompt box and per-clip pricing. No apps, no API keys, no subscriptions. Pick a model to see what it's best at.
Free credits to start · no card · switch models any time
Newest Seedance: top quality and motion, clips up to 30s with synced audio.
- Text & image to video
- Clips up to 30s
- Synced audio & dialogue
Great quality and motion at a balanced price, with synced audio.
- Text & image to video
- Clips up to 15s
- Synced audio & dialogue
MiniMax's flagship: cinematic motion with native audio, up to 2K, clips up to 15s.
- Text & image to video
- Clips up to 15s
- No content filters (18+ gate)
Newest Wan: one continuous take up to 30s, with audio. Unfiltered.
- Text & image to video
- Clips up to 30s
- No content filters (18+ gate)
Lifelike people and faces, with audio. Clips up to 15s.
- Text & image to video
- Clips up to 15s
- No content filters (18+ gate)
Sharp detail and fluid motion with audio, up to 15s. Filtered.
- Text & image to video
- Clips up to 15s
- Audio with video
by Gemini · Latest, highest quality
~89 credits per image
Try now
by OpenAI · OpenAI's fastest high-fidelity image model
67 credits per image
Try now
Qwen Image 3.0 ProUncensored
by Alibaba · Top-quality unrestricted images at a great price
from 53 credits per image
Try now
Seedream 5.0Uncensored
by ByteDance · Newest flagship Seedream — highest quality
from 60 credits per image
Try nowSeedream 5.0 LiteUncensored
by ByteDance · Newest Seedream — fast, high-resolution images
47 credits per image
Try nowSeedream 4.5Uncensored
by ByteDance · High-resolution cinematic image generation
67 credits per image
Try now
by xAI · xAI's newest, higher-quality image generation
53 credits per image
Try now
All of them live in one studio — generate, compare and edit with the same prompt box.
Try nowHaiku 4.5
by Anthropic · Fast, low-cost replies for everyday tasks
- Try now
Sonnet 5.5
by Anthropic · Balanced all-rounder for writing and code
- Try now
Opus 5.5
by Anthropic · Latest Opus — top-tier reasoning for complex work
- Try now
Fable 5.1
by Anthropic · Anthropic's most capable model for demanding work
- Try now
GPT-6 Astra
by OpenAI · OpenAI's flagship model for agentic and computer-use work
- Try now
GPT-4.1 Nano
by OpenAI · Cheapest option for simple tasks
- Try now
GPT-5 Mini
by OpenAI · Fast and affordable for daily use
- Try now
GPT-6 Luna
by OpenAI · Fast, low-cost GPT for high-volume everyday work
- Try now
GPT-5.6 Terra
by OpenAI · Latest GPT model with strong reasoning at balanced cost
- Try now
GPT-6 Sol
by OpenAI · Most capable, cost-efficient GPT-6 model
- Try now
Gemini 2.5 Flash Lite
by Google · Cheapest for high-volume tasks
- Try now
Gemini 2.5 Pro
by Google · Deep reasoning with a huge context window
- Try now
Gemini 3.8 Flash
by Google · Latest Gemini Flash for complex agentic and multimodal work
- Try now
Gemini 3.6 Flash
by Google · Latest Gemini for agentic and multimodal work
- Try now
Gemini 3.5 Flash Lite
by Google · Low-cost Gemini for fast, high-volume tasks
- Try now
Grok 4.7
by xAI · Latest Grok with strong reasoning and live web awareness
- Try now
Grok 4.3
by xAI · Top-tier reasoning and capability at a lower price
Try nowDeepSeek V4 Flash
by DeepSeek · Fast, low-cost DeepSeek model
- Try now
DeepSeek V4 ProUncensored
by DeepSeek · Deep reasoning model with a 1024K context window
- Try now
Seed 2.0 Lite
by ByteDance · Lightweight and budget-friendly
- Try now
Seed 1.8
by ByteDance · Most capable Seed model, strong multilingual
Try nowHermes 4 405BUncensored
by Nous Research · Flagship Hermes — strong reasoning, steerable and uncensored
Try nowHy3
by Tencent · Low-cost open-weight model for reasoning and agentic work
- Try now
Hy4 Preview
by Tencent · Tencent's open-weight coding/agentic model with a 1M-token context window
Try nowGLM-5.2
by Z.ai · Z.ai flagship reasoning model with lower cost than Opus
Every one of them behind the same private prompt box — switch mid-conversation any time.
Not sure which model? Just start.
Every model here shares one prompt box — run the same idea on two or three of them for cents and keep the winner.