Changes

Qwen-VL-Plus and Qwen-VL-Max will support implicit caching. Implicit cache is the default automatic mode with no configuration required, suitable for general scenarios. The system automatically detects common prefixes in request content and caches them; cached tokens are billed at 20% of the input price.

Impacts

After the update:

Cached price for Qwen-VL-Plus and Qwen-VL-Max in the Singapore region:

Thank you for your continued support in Model Studio.