WaveSpeedAI Announces Partnership with Alibaba for the Global Launch of Wan 2.5
Singapore, September 26, 2025
WaveSpeedAI, a market-leading multimodal model API aggregation platform, has announced the partnership today with Alibaba. As the first third-party inference acceleration platform to integrate and commercially launch the Wan 2.5 Model. WaveSpeedAI now supports both image and video generation tasks via API for global users.
This partnership extends beyond model integration, with WaveSpeedAI leveraging Alibaba Cloud’s elastic heterogeneous resources and co-optimizing deployment and scheduling to ensure stable, low-latency inference for large-scale Wan 2.5 deployments.
Key Advancements of Wan 2.5 on WaveSpeedAI:
- Cost efficiency: Up to 3× cheaper than Google Veo 3, democratizing high-quality video generation.
- End-to-end A/V sync: Single-pass generation with native lip-sync, sound effects, and background audio.
- Multilingual support: Beyond English, Wan 2.5 natively supports Chinese and minor languages.
- Flexible output: Videos up to 10 seconds, with multiple aspect ratios for different publishing needs.
- Voice-driven generation: Direct input of speech, music, or effects to guide video creation.
Through its partnership with Alibaba, WaveSpeedAI powers commercial use cases for clients such as Freepik and Higgsfield, further expanding the frontiers of multimodal AI infrastructure and empowering creators worldwide.
👉 Experience Wan 2.5 now via the WaveSpeedAI API.
Stay connected with us:
© 2025 WaveSpeedAI. All rights reserved.