Model APIs Updates
New Features
-
New Model Available: Qwen-Image
The Qwen-Image is a 20B MMDIT model for next-gen text-to-image generation. Especially strong at creating stunning graphic posters with native text.
Key Highlights:
- SOTA text rendering — rivals GPT-4o in English, best-in-class for Chinese
- In-pixel text generation — no overlays, fully integrated
- Bilingual support, diverse fonts, complex layouts
-
New Model Available: Wan 2.2 Text to Video, Wan 2.2 Image to Video
Wan2.2 is the world’s first open-source MoE-architecture video generation model with cinematic control. Its Mixture-of-Experts (MoE) architecture scales model capacity without increasing computational requirements.
Key Highlights:
- First-of-its-kind MoE Architecture: The first model to apply an MoE framework to video generation, drastically cutting down on resource consumption.
- 50% More Efficient: Saves approximately 50% of computational resources compared to models of a similar parameter scale.
- Enhanced Capabilities: Achieves significant improvements in generating complex motion, character interactions, and overall aesthetic quality.
GPUs Updates
New Features
- Add region selection when creating an endpoint in Serverless Novita now supports creating endpoints in selected regions. Updated time: August 5, 2025