Skip to main content
2026/9/24 23:12 (UTC+8) · Service Notice · Alibaba ✅ Alibaba Cloud Model Studio routing optimized: image requests are noticeably faster We have optimized our cross-border network path to Alibaba Cloud Model Studio, so models served through Alibaba Cloud now respond faster and more steadily. Multimodal requests with images benefit the most, especially image understanding calls that send images as base64, where the wait for the first token is noticeably shorter. Covered models:
  • qwen3.8-max, qwen3.8-flash and other Qwen text and multimodal models
  • DeepSeek, GLM, Kimi and other models served through Alibaba Cloud
  • Wan and HappyHorse video models, and Qwen Omni Realtime models
No code or configuration changes are needed. If you still see noticeably slow responses, contact support with the request time and we will look into each one.
← Back to Live Updates · 📚 Monthly archive